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"
85 This will try to automatically determine if the current tree is a
86 release tree by looking for git tags that belong to the current
89 A string of the format -gxxxxxxxx will be added to the localversion
90 if a git-based tree is found. The string generated by this will be
91 appended after any matching localversion* files, and after the value
92 set in CONFIG_LOCALVERSION.
94 (The actual string used here is the first eight characters produced
95 by running the command:
97 $ git rev-parse --verify HEAD
99 which is done within the script "scripts/setlocalversion".)
101 config HAVE_KERNEL_GZIP
104 config HAVE_KERNEL_BZIP2
107 config HAVE_KERNEL_LZMA
110 config HAVE_KERNEL_XZ
113 config HAVE_KERNEL_LZO
116 config HAVE_KERNEL_LZ4
120 prompt "Kernel compression mode"
122 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4
124 The linux kernel is a kind of self-extracting executable.
125 Several compression algorithms are available, which differ
126 in efficiency, compression and decompression speed.
127 Compression speed is only relevant when building a kernel.
128 Decompression speed is relevant at each boot.
130 If you have any problems with bzip2 or lzma compressed
131 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
132 version of this functionality (bzip2 only), for 2.4, was
133 supplied by Christian Ludwig)
135 High compression options are mostly useful for users, who
136 are low on disk space (embedded systems), but for whom ram
139 If in doubt, select 'gzip'
143 depends on HAVE_KERNEL_GZIP
145 The old and tried gzip compression. It provides a good balance
146 between compression ratio and decompression speed.
150 depends on HAVE_KERNEL_BZIP2
152 Its compression ratio and speed is intermediate.
153 Decompression speed is slowest among the choices. The kernel
154 size is about 10% smaller with bzip2, in comparison to gzip.
155 Bzip2 uses a large amount of memory. For modern kernels you
156 will need at least 8MB RAM or more for booting.
160 depends on HAVE_KERNEL_LZMA
162 This compression algorithm's ratio is best. Decompression speed
163 is between gzip and bzip2. Compression is slowest.
164 The kernel size is about 33% smaller with LZMA in comparison to gzip.
168 depends on HAVE_KERNEL_XZ
170 XZ uses the LZMA2 algorithm and instruction set specific
171 BCJ filters which can improve compression ratio of executable
172 code. The size of the kernel is about 30% smaller with XZ in
173 comparison to gzip. On architectures for which there is a BCJ
174 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
175 will create a few percent smaller kernel than plain LZMA.
177 The speed is about the same as with LZMA: The decompression
178 speed of XZ is better than that of bzip2 but worse than gzip
179 and LZO. Compression is slow.
183 depends on HAVE_KERNEL_LZO
185 Its compression ratio is the poorest among the choices. The kernel
186 size is about 10% bigger than gzip; however its speed
187 (both compression and decompression) is the fastest.
191 depends on HAVE_KERNEL_LZ4
193 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
194 A preliminary version of LZ4 de/compression tool is available at
195 <https://code.google.com/p/lz4/>.
197 Its compression ratio is worse than LZO. The size of the kernel
198 is about 8% bigger than LZO. But the decompression speed is
203 config DEFAULT_HOSTNAME
204 string "Default hostname"
207 This option determines the default system hostname before userspace
208 calls sethostname(2). The kernel traditionally uses "(none)" here,
209 but you may wish to use a different default here to make a minimal
210 system more usable with less configuration.
213 bool "Support for paging of anonymous memory (swap)"
214 depends on MMU && BLOCK
217 This option allows you to choose whether you want to have support
218 for so called swap devices or swap files in your kernel that are
219 used to provide more virtual memory than the actual RAM present
220 in your computer. If unsure say Y.
225 Inter Process Communication is a suite of library functions and
226 system calls which let processes (running programs) synchronize and
227 exchange information. It is generally considered to be a good thing,
228 and some programs won't run unless you say Y here. In particular, if
229 you want to run the DOS emulator dosemu under Linux (read the
230 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
231 you'll need to say Y here.
233 You can find documentation about IPC with "info ipc" and also in
234 section 6.4 of the Linux Programmer's Guide, available from
235 <http://www.tldp.org/guides.html>.
237 config SYSVIPC_SYSCTL
244 bool "POSIX Message Queues"
247 POSIX variant of message queues is a part of IPC. In POSIX message
248 queues every message has a priority which decides about succession
249 of receiving it by a process. If you want to compile and run
250 programs written e.g. for Solaris with use of its POSIX message
251 queues (functions mq_*) say Y here.
253 POSIX message queues are visible as a filesystem called 'mqueue'
254 and can be mounted somewhere if you want to do filesystem
255 operations on message queues.
259 config POSIX_MQUEUE_SYSCTL
261 depends on POSIX_MQUEUE
265 config CROSS_MEMORY_ATTACH
266 bool "Enable process_vm_readv/writev syscalls"
270 Enabling this option adds the system calls process_vm_readv and
271 process_vm_writev which allow a process with the correct privileges
272 to directly read from or write to another process' address space.
273 See the man page for more details.
276 bool "open by fhandle syscalls" if EXPERT
280 If you say Y here, a user level program will be able to map
281 file names to handle and then later use the handle for
282 different file system operations. This is useful in implementing
283 userspace file servers, which now track files using handles instead
284 of names. The handle would remain the same even if file names
285 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
289 bool "uselib syscall"
290 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
292 This option enables the uselib syscall, a system call used in the
293 dynamic linker from libc5 and earlier. glibc does not use this
294 system call. If you intend to run programs built on libc5 or
295 earlier, you may need to enable this syscall. Current systems
296 running glibc can safely disable this.
299 bool "Auditing support"
302 Enable auditing infrastructure that can be used with another
303 kernel subsystem, such as SELinux (which requires this for
304 logging of avc messages output). System call auditing is included
305 on architectures which support it.
307 config HAVE_ARCH_AUDITSYSCALL
312 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
316 depends on AUDITSYSCALL
321 depends on AUDITSYSCALL
324 source "kernel/irq/Kconfig"
325 source "kernel/time/Kconfig"
327 menu "CPU/Task time and stats accounting"
329 config VIRT_CPU_ACCOUNTING
333 prompt "Cputime accounting"
334 default TICK_CPU_ACCOUNTING if !PPC64
335 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
337 # Kind of a stub config for the pure tick based cputime accounting
338 config TICK_CPU_ACCOUNTING
339 bool "Simple tick based cputime accounting"
340 depends on !S390 && !NO_HZ_FULL
342 This is the basic tick based cputime accounting that maintains
343 statistics about user, system and idle time spent on per jiffies
348 config VIRT_CPU_ACCOUNTING_NATIVE
349 bool "Deterministic task and CPU time accounting"
350 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
351 select VIRT_CPU_ACCOUNTING
353 Select this option to enable more accurate task and CPU time
354 accounting. This is done by reading a CPU counter on each
355 kernel entry and exit and on transitions within the kernel
356 between system, softirq and hardirq state, so there is a
357 small performance impact. In the case of s390 or IBM POWER > 5,
358 this also enables accounting of stolen time on logically-partitioned
361 config VIRT_CPU_ACCOUNTING_GEN
362 bool "Full dynticks CPU time accounting"
363 depends on HAVE_CONTEXT_TRACKING
364 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
365 select VIRT_CPU_ACCOUNTING
366 select CONTEXT_TRACKING
368 Select this option to enable task and CPU time accounting on full
369 dynticks systems. This accounting is implemented by watching every
370 kernel-user boundaries using the context tracking subsystem.
371 The accounting is thus performed at the expense of some significant
374 For now this is only useful if you are working on the full
375 dynticks subsystem development.
381 config IRQ_TIME_ACCOUNTING
382 bool "Fine granularity task level IRQ time accounting"
383 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
385 Select this option to enable fine granularity task irq time
386 accounting. This is done by reading a timestamp on each
387 transitions between softirq and hardirq state, so there can be a
388 small performance impact.
390 If in doubt, say N here.
392 config BSD_PROCESS_ACCT
393 bool "BSD Process Accounting"
396 If you say Y here, a user level program will be able to instruct the
397 kernel (via a special system call) to write process accounting
398 information to a file: whenever a process exits, information about
399 that process will be appended to the file by the kernel. The
400 information includes things such as creation time, owning user,
401 command name, memory usage, controlling terminal etc. (the complete
402 list is in the struct acct in <file:include/linux/acct.h>). It is
403 up to the user level program to do useful things with this
404 information. This is generally a good idea, so say Y.
406 config BSD_PROCESS_ACCT_V3
407 bool "BSD Process Accounting version 3 file format"
408 depends on BSD_PROCESS_ACCT
411 If you say Y here, the process accounting information is written
412 in a new file format that also logs the process IDs of each
413 process and it's parent. Note that this file format is incompatible
414 with previous v0/v1/v2 file formats, so you will need updated tools
415 for processing it. A preliminary version of these tools is available
416 at <http://www.gnu.org/software/acct/>.
419 bool "Export task/process statistics through netlink"
424 Export selected statistics for tasks/processes through the
425 generic netlink interface. Unlike BSD process accounting, the
426 statistics are available during the lifetime of tasks/processes as
427 responses to commands. Like BSD accounting, they are sent to user
432 config TASK_DELAY_ACCT
433 bool "Enable per-task delay accounting"
437 Collect information on time spent by a task waiting for system
438 resources like cpu, synchronous block I/O completion and swapping
439 in pages. Such statistics can help in setting a task's priorities
440 relative to other tasks for cpu, io, rss limits etc.
445 bool "Enable extended accounting over taskstats"
448 Collect extended task accounting data and send the data
449 to userland for processing over the taskstats interface.
453 config TASK_IO_ACCOUNTING
454 bool "Enable per-task storage I/O accounting"
455 depends on TASK_XACCT
457 Collect information on the number of bytes of storage I/O which this
462 endmenu # "CPU/Task time and stats accounting"
468 default y if !PREEMPT && SMP
470 This option selects the RCU implementation that is
471 designed for very large SMP system with hundreds or
472 thousands of CPUs. It also scales down nicely to
479 This option selects the RCU implementation that is
480 designed for very large SMP systems with hundreds or
481 thousands of CPUs, but for which real-time response
482 is also required. It also scales down nicely to
485 Select this option if you are unsure.
489 default y if !PREEMPT && !SMP
491 This option selects the RCU implementation that is
492 designed for UP systems from which real-time response
493 is not required. This option greatly reduces the
494 memory footprint of RCU.
497 bool "Make expert-level adjustments to RCU configuration"
500 This option needs to be enabled if you wish to make
501 expert-level adjustments to RCU configuration. By default,
502 no such adjustments can be made, which has the often-beneficial
503 side-effect of preventing "make oldconfig" from asking you all
504 sorts of detailed questions about how you would like numerous
505 obscure RCU options to be set up.
507 Say Y if you need to make expert-level adjustments to RCU.
509 Say N if you are unsure.
514 This option selects the sleepable version of RCU. This version
515 permits arbitrary sleeping or blocking within RCU read-side critical
524 This option enables a task-based RCU implementation that uses
525 only voluntary context switch (not preemption!), idle, and
526 user-mode execution as quiescent states.
528 config RCU_STALL_COMMON
529 def_bool ( TREE_RCU || PREEMPT_RCU || RCU_TRACE )
531 This option enables RCU CPU stall code that is common between
532 the TINY and TREE variants of RCU. The purpose is to allow
533 the tiny variants to disable RCU CPU stall warnings, while
534 making these warnings mandatory for the tree variants.
536 config CONTEXT_TRACKING
539 config CONTEXT_TRACKING_FORCE
540 bool "Force context tracking"
541 depends on CONTEXT_TRACKING
542 default y if !NO_HZ_FULL
544 The major pre-requirement for full dynticks to work is to
545 support the context tracking subsystem. But there are also
546 other dependencies to provide in order to make the full
549 This option stands for testing when an arch implements the
550 context tracking backend but doesn't yet fullfill all the
551 requirements to make the full dynticks feature working.
552 Without the full dynticks, there is no way to test the support
553 for context tracking and the subsystems that rely on it: RCU
554 userspace extended quiescent state and tickless cputime
555 accounting. This option copes with the absence of the full
556 dynticks subsystem by forcing the context tracking on all
559 Say Y only if you're working on the development of an
560 architecture backend for the context tracking.
562 Say N otherwise, this option brings an overhead that you
563 don't want in production.
567 int "Tree-based hierarchical RCU fanout value"
570 depends on (TREE_RCU || PREEMPT_RCU) && RCU_EXPERT
574 This option controls the fanout of hierarchical implementations
575 of RCU, allowing RCU to work efficiently on machines with
576 large numbers of CPUs. This value must be at least the fourth
577 root of NR_CPUS, which allows NR_CPUS to be insanely large.
578 The default value of RCU_FANOUT should be used for production
579 systems, but if you are stress-testing the RCU implementation
580 itself, small RCU_FANOUT values allow you to test large-system
581 code paths on small(er) systems.
583 Select a specific number if testing RCU itself.
584 Take the default if unsure.
586 config RCU_FANOUT_LEAF
587 int "Tree-based hierarchical RCU leaf-level fanout value"
590 depends on (TREE_RCU || PREEMPT_RCU) && RCU_EXPERT
593 This option controls the leaf-level fanout of hierarchical
594 implementations of RCU, and allows trading off cache misses
595 against lock contention. Systems that synchronize their
596 scheduling-clock interrupts for energy-efficiency reasons will
597 want the default because the smaller leaf-level fanout keeps
598 lock contention levels acceptably low. Very large systems
599 (hundreds or thousands of CPUs) will instead want to set this
600 value to the maximum value possible in order to reduce the
601 number of cache misses incurred during RCU's grace-period
602 initialization. These systems tend to run CPU-bound, and thus
603 are not helped by synchronized interrupts, and thus tend to
604 skew them, which reduces lock contention enough that large
605 leaf-level fanouts work well.
607 Select a specific number if testing RCU itself.
609 Select the maximum permissible value for large systems.
611 Take the default if unsure.
613 config RCU_FAST_NO_HZ
614 bool "Accelerate last non-dyntick-idle CPU's grace periods"
615 depends on NO_HZ_COMMON && SMP && RCU_EXPERT
618 This option permits CPUs to enter dynticks-idle state even if
619 they have RCU callbacks queued, and prevents RCU from waking
620 these CPUs up more than roughly once every four jiffies (by
621 default, you can adjust this using the rcutree.rcu_idle_gp_delay
622 parameter), thus improving energy efficiency. On the other
623 hand, this option increases the duration of RCU grace periods,
624 for example, slowing down synchronize_rcu().
626 Say Y if energy efficiency is critically important, and you
627 don't care about increased grace-period durations.
629 Say N if you are unsure.
631 config TREE_RCU_TRACE
632 def_bool RCU_TRACE && ( TREE_RCU || PREEMPT_RCU )
635 This option provides tracing for the TREE_RCU and
636 PREEMPT_RCU implementations, permitting Makefile to
637 trivially select kernel/rcutree_trace.c.
640 bool "Enable RCU priority boosting"
641 depends on RT_MUTEXES && PREEMPT_RCU && RCU_EXPERT
644 This option boosts the priority of preempted RCU readers that
645 block the current preemptible RCU grace period for too long.
646 This option also prevents heavy loads from blocking RCU
647 callback invocation for all flavors of RCU.
649 Say Y here if you are working with real-time apps or heavy loads
650 Say N here if you are unsure.
652 config RCU_KTHREAD_PRIO
653 int "Real-time priority to use for RCU worker threads"
654 range 1 99 if RCU_BOOST
655 range 0 99 if !RCU_BOOST
656 default 1 if RCU_BOOST
657 default 0 if !RCU_BOOST
658 depends on RCU_EXPERT
660 This option specifies the SCHED_FIFO priority value that will be
661 assigned to the rcuc/n and rcub/n threads and is also the value
662 used for RCU_BOOST (if enabled). If you are working with a
663 real-time application that has one or more CPU-bound threads
664 running at a real-time priority level, you should set
665 RCU_KTHREAD_PRIO to a priority higher than the highest-priority
666 real-time CPU-bound application thread. The default RCU_KTHREAD_PRIO
667 value of 1 is appropriate in the common case, which is real-time
668 applications that do not have any CPU-bound threads.
670 Some real-time applications might not have a single real-time
671 thread that saturates a given CPU, but instead might have
672 multiple real-time threads that, taken together, fully utilize
673 that CPU. In this case, you should set RCU_KTHREAD_PRIO to
674 a priority higher than the lowest-priority thread that is
675 conspiring to prevent the CPU from running any non-real-time
676 tasks. For example, if one thread at priority 10 and another
677 thread at priority 5 are between themselves fully consuming
678 the CPU time on a given CPU, then RCU_KTHREAD_PRIO should be
679 set to priority 6 or higher.
681 Specify the real-time priority, or take the default if unsure.
683 config RCU_BOOST_DELAY
684 int "Milliseconds to delay boosting after RCU grace-period start"
689 This option specifies the time to wait after the beginning of
690 a given grace period before priority-boosting preempted RCU
691 readers blocking that grace period. Note that any RCU reader
692 blocking an expedited RCU grace period is boosted immediately.
694 Accept the default if unsure.
697 bool "Offload RCU callback processing from boot-selected CPUs"
698 depends on TREE_RCU || PREEMPT_RCU
699 depends on RCU_EXPERT || NO_HZ_FULL
702 Use this option to reduce OS jitter for aggressive HPC or
703 real-time workloads. It can also be used to offload RCU
704 callback invocation to energy-efficient CPUs in battery-powered
705 asymmetric multiprocessors.
707 This option offloads callback invocation from the set of
708 CPUs specified at boot time by the rcu_nocbs parameter.
709 For each such CPU, a kthread ("rcuox/N") will be created to
710 invoke callbacks, where the "N" is the CPU being offloaded,
711 and where the "x" is "b" for RCU-bh, "p" for RCU-preempt, and
712 "s" for RCU-sched. Nothing prevents this kthread from running
713 on the specified CPUs, but (1) the kthreads may be preempted
714 between each callback, and (2) affinity or cgroups can be used
715 to force the kthreads to run on whatever set of CPUs is desired.
717 Say Y here if you want to help to debug reduced OS jitter.
718 Say N here if you are unsure.
721 prompt "Build-forced no-CBs CPUs"
722 default RCU_NOCB_CPU_NONE
723 depends on RCU_NOCB_CPU
725 This option allows no-CBs CPUs (whose RCU callbacks are invoked
726 from kthreads rather than from softirq context) to be specified
727 at build time. Additional no-CBs CPUs may be specified by
728 the rcu_nocbs= boot parameter.
730 config RCU_NOCB_CPU_NONE
731 bool "No build_forced no-CBs CPUs"
733 This option does not force any of the CPUs to be no-CBs CPUs.
734 Only CPUs designated by the rcu_nocbs= boot parameter will be
735 no-CBs CPUs, whose RCU callbacks will be invoked by per-CPU
736 kthreads whose names begin with "rcuo". All other CPUs will
737 invoke their own RCU callbacks in softirq context.
739 Select this option if you want to choose no-CBs CPUs at
740 boot time, for example, to allow testing of different no-CBs
741 configurations without having to rebuild the kernel each time.
743 config RCU_NOCB_CPU_ZERO
744 bool "CPU 0 is a build_forced no-CBs CPU"
746 This option forces CPU 0 to be a no-CBs CPU, so that its RCU
747 callbacks are invoked by a per-CPU kthread whose name begins
748 with "rcuo". Additional CPUs may be designated as no-CBs
749 CPUs using the rcu_nocbs= boot parameter will be no-CBs CPUs.
750 All other CPUs will invoke their own RCU callbacks in softirq
753 Select this if CPU 0 needs to be a no-CBs CPU for real-time
754 or energy-efficiency reasons, but the real reason it exists
755 is to ensure that randconfig testing covers mixed systems.
757 config RCU_NOCB_CPU_ALL
758 bool "All CPUs are build_forced no-CBs CPUs"
760 This option forces all CPUs to be no-CBs CPUs. The rcu_nocbs=
761 boot parameter will be ignored. All CPUs' RCU callbacks will
762 be executed in the context of per-CPU rcuo kthreads created for
763 this purpose. Assuming that the kthreads whose names start with
764 "rcuo" are bound to "housekeeping" CPUs, this reduces OS jitter
765 on the remaining CPUs, but might decrease memory locality during
766 RCU-callback invocation, thus potentially degrading throughput.
768 Select this if all CPUs need to be no-CBs CPUs for real-time
769 or energy-efficiency reasons.
773 config RCU_EXPEDITE_BOOT
777 This option enables expedited grace periods at boot time,
778 as if rcu_expedite_gp() had been invoked early in boot.
779 The corresponding rcu_unexpedite_gp() is invoked from
780 rcu_end_inkernel_boot(), which is intended to be invoked
781 at the end of the kernel-only boot sequence, just before
784 Accept the default if unsure.
786 endmenu # "RCU Subsystem"
793 tristate "Kernel .config support"
796 This option enables the complete Linux kernel ".config" file
797 contents to be saved in the kernel. It provides documentation
798 of which kernel options are used in a running kernel or in an
799 on-disk kernel. This information can be extracted from the kernel
800 image file with the script scripts/extract-ikconfig and used as
801 input to rebuild the current kernel or to build another kernel.
802 It can also be extracted from a running kernel by reading
803 /proc/config.gz if enabled (below).
806 bool "Enable access to .config through /proc/config.gz"
807 depends on IKCONFIG && PROC_FS
809 This option enables access to the kernel configuration file
810 through /proc/config.gz.
813 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
818 Select the minimal kernel log buffer size as a power of 2.
819 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
820 parameter, see below. Any higher size also might be forced
821 by "log_buf_len" boot parameter.
831 config LOG_CPU_MAX_BUF_SHIFT
832 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
835 default 12 if !BASE_SMALL
836 default 0 if BASE_SMALL
839 This option allows to increase the default ring buffer size
840 according to the number of CPUs. The value defines the contribution
841 of each CPU as a power of 2. The used space is typically only few
842 lines however it might be much more when problems are reported,
845 The increased size means that a new buffer has to be allocated and
846 the original static one is unused. It makes sense only on systems
847 with more CPUs. Therefore this value is used only when the sum of
848 contributions is greater than the half of the default kernel ring
849 buffer as defined by LOG_BUF_SHIFT. The default values are set
850 so that more than 64 CPUs are needed to trigger the allocation.
852 Also this option is ignored when "log_buf_len" kernel parameter is
853 used as it forces an exact (power of two) size of the ring buffer.
855 The number of possible CPUs is used for this computation ignoring
856 hotplugging making the computation optimal for the worst case
857 scenario while allowing a simple algorithm to be used from bootup.
859 Examples shift values and their meaning:
860 17 => 128 KB for each CPU
861 16 => 64 KB for each CPU
862 15 => 32 KB for each CPU
863 14 => 16 KB for each CPU
864 13 => 8 KB for each CPU
865 12 => 4 KB for each CPU
867 config NMI_LOG_BUF_SHIFT
868 int "Temporary per-CPU NMI log buffer size (12 => 4KB, 13 => 8KB)"
871 depends on PRINTK_NMI
873 Select the size of a per-CPU buffer where NMI messages are temporary
874 stored. They are copied to the main log buffer in a safe context
875 to avoid a deadlock. The value defines the size as a power of 2.
877 NMI messages are rare and limited. The largest one is when
878 a backtrace is printed. It usually fits into 4KB. Select
879 8KB if you want to be on the safe side.
882 17 => 128 KB for each CPU
883 16 => 64 KB for each CPU
884 15 => 32 KB for each CPU
885 14 => 16 KB for each CPU
886 13 => 8 KB for each CPU
887 12 => 4 KB for each CPU
890 # Architectures with an unreliable sched_clock() should select this:
892 config HAVE_UNSTABLE_SCHED_CLOCK
895 config GENERIC_SCHED_CLOCK
899 # For architectures that want to enable the support for NUMA-affine scheduler
902 config ARCH_SUPPORTS_NUMA_BALANCING
906 # For architectures that prefer to flush all TLBs after a number of pages
907 # are unmapped instead of sending one IPI per page to flush. The architecture
908 # must provide guarantees on what happens if a clean TLB cache entry is
909 # written after the unmap. Details are in mm/rmap.c near the check for
910 # should_defer_flush. The architecture should also consider if the full flush
911 # and the refill costs are offset by the savings of sending fewer IPIs.
912 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
916 # For architectures that know their GCC __int128 support is sound
918 config ARCH_SUPPORTS_INT128
921 # For architectures that (ab)use NUMA to represent different memory regions
922 # all cpu-local but of different latencies, such as SuperH.
924 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
927 config NUMA_BALANCING
928 bool "Memory placement aware NUMA scheduler"
929 depends on ARCH_SUPPORTS_NUMA_BALANCING
930 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
931 depends on SMP && NUMA && MIGRATION
933 This option adds support for automatic NUMA aware memory/task placement.
934 The mechanism is quite primitive and is based on migrating memory when
935 it has references to the node the task is running on.
937 This system will be inactive on UMA systems.
939 config NUMA_BALANCING_DEFAULT_ENABLED
940 bool "Automatically enable NUMA aware memory/task placement"
942 depends on NUMA_BALANCING
944 If set, automatic NUMA balancing will be enabled if running on a NUMA
948 bool "Control Group support"
951 This option adds support for grouping sets of processes together, for
952 use with process control subsystems such as Cpusets, CFS, memory
953 controls or device isolation.
955 - Documentation/scheduler/sched-design-CFS.txt (CFS)
956 - Documentation/cgroup-v1/ (features for grouping, isolation
957 and resource control)
967 bool "Memory controller"
971 Provides control over the memory footprint of tasks in a cgroup.
974 bool "Swap controller"
975 depends on MEMCG && SWAP
977 Provides control over the swap space consumed by tasks in a cgroup.
979 config MEMCG_SWAP_ENABLED
980 bool "Swap controller enabled by default"
981 depends on MEMCG_SWAP
984 Memory Resource Controller Swap Extension comes with its price in
985 a bigger memory consumption. General purpose distribution kernels
986 which want to enable the feature but keep it disabled by default
987 and let the user enable it by swapaccount=1 boot command line
988 parameter should have this option unselected.
989 For those who want to have the feature enabled by default should
990 select this option (if, for some reason, they need to disable it
991 then swapaccount=0 does the trick).
998 Generic block IO controller cgroup interface. This is the common
999 cgroup interface which should be used by various IO controlling
1002 Currently, CFQ IO scheduler uses it to recognize task groups and
1003 control disk bandwidth allocation (proportional time slice allocation)
1004 to such task groups. It is also used by bio throttling logic in
1005 block layer to implement upper limit in IO rates on a device.
1007 This option only enables generic Block IO controller infrastructure.
1008 One needs to also enable actual IO controlling logic/policy. For
1009 enabling proportional weight division of disk bandwidth in CFQ, set
1010 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
1011 CONFIG_BLK_DEV_THROTTLING=y.
1013 See Documentation/cgroup-v1/blkio-controller.txt for more information.
1015 config DEBUG_BLK_CGROUP
1016 bool "IO controller debugging"
1017 depends on BLK_CGROUP
1020 Enable some debugging help. Currently it exports additional stat
1021 files in a cgroup which can be useful for debugging.
1023 config CGROUP_WRITEBACK
1025 depends on MEMCG && BLK_CGROUP
1028 menuconfig CGROUP_SCHED
1029 bool "CPU controller"
1032 This feature lets CPU scheduler recognize task groups and control CPU
1033 bandwidth allocation to such task groups. It uses cgroups to group
1037 config FAIR_GROUP_SCHED
1038 bool "Group scheduling for SCHED_OTHER"
1039 depends on CGROUP_SCHED
1040 default CGROUP_SCHED
1042 config CFS_BANDWIDTH
1043 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
1044 depends on FAIR_GROUP_SCHED
1047 This option allows users to define CPU bandwidth rates (limits) for
1048 tasks running within the fair group scheduler. Groups with no limit
1049 set are considered to be unconstrained and will run with no
1051 See tip/Documentation/scheduler/sched-bwc.txt for more information.
1053 config RT_GROUP_SCHED
1054 bool "Group scheduling for SCHED_RR/FIFO"
1055 depends on CGROUP_SCHED
1058 This feature lets you explicitly allocate real CPU bandwidth
1059 to task groups. If enabled, it will also make it impossible to
1060 schedule realtime tasks for non-root users until you allocate
1061 realtime bandwidth for them.
1062 See Documentation/scheduler/sched-rt-group.txt for more information.
1067 bool "PIDs controller"
1069 Provides enforcement of process number limits in the scope of a
1070 cgroup. Any attempt to fork more processes than is allowed in the
1071 cgroup will fail. PIDs are fundamentally a global resource because it
1072 is fairly trivial to reach PID exhaustion before you reach even a
1073 conservative kmemcg limit. As a result, it is possible to grind a
1074 system to halt without being limited by other cgroup policies. The
1075 PIDs controller is designed to stop this from happening.
1077 It should be noted that organisational operations (such as attaching
1078 to a cgroup hierarchy will *not* be blocked by the PIDs controller),
1079 since the PIDs limit only affects a process's ability to fork, not to
1082 config CGROUP_FREEZER
1083 bool "Freezer controller"
1085 Provides a way to freeze and unfreeze all tasks in a
1088 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
1089 controller includes important in-kernel memory consumers per default.
1091 If you're using cgroup2, say N.
1093 config CGROUP_HUGETLB
1094 bool "HugeTLB controller"
1095 depends on HUGETLB_PAGE
1099 Provides a cgroup controller for HugeTLB pages.
1100 When you enable this, you can put a per cgroup limit on HugeTLB usage.
1101 The limit is enforced during page fault. Since HugeTLB doesn't
1102 support page reclaim, enforcing the limit at page fault time implies
1103 that, the application will get SIGBUS signal if it tries to access
1104 HugeTLB pages beyond its limit. This requires the application to know
1105 beforehand how much HugeTLB pages it would require for its use. The
1106 control group is tracked in the third page lru pointer. This means
1107 that we cannot use the controller with huge page less than 3 pages.
1110 bool "Cpuset controller"
1112 This option will let you create and manage CPUSETs which
1113 allow dynamically partitioning a system into sets of CPUs and
1114 Memory Nodes and assigning tasks to run only within those sets.
1115 This is primarily useful on large SMP or NUMA systems.
1119 config PROC_PID_CPUSET
1120 bool "Include legacy /proc/<pid>/cpuset file"
1124 config CGROUP_DEVICE
1125 bool "Device controller"
1127 Provides a cgroup controller implementing whitelists for
1128 devices which a process in the cgroup can mknod or open.
1130 config CGROUP_CPUACCT
1131 bool "Simple CPU accounting controller"
1133 Provides a simple controller for monitoring the
1134 total CPU consumed by the tasks in a cgroup.
1137 bool "Perf controller"
1138 depends on PERF_EVENTS
1140 This option extends the perf per-cpu mode to restrict monitoring
1141 to threads which belong to the cgroup specified and run on the
1147 bool "Example controller"
1150 This option enables a simple controller that exports
1151 debugging information about the cgroups framework.
1157 config CHECKPOINT_RESTORE
1158 bool "Checkpoint/restore support" if EXPERT
1159 select PROC_CHILDREN
1162 Enables additional kernel features in a sake of checkpoint/restore.
1163 In particular it adds auxiliary prctl codes to setup process text,
1164 data and heap segment sizes, and a few additional /proc filesystem
1167 If unsure, say N here.
1169 menuconfig NAMESPACES
1170 bool "Namespaces support" if EXPERT
1171 depends on MULTIUSER
1174 Provides the way to make tasks work with different objects using
1175 the same id. For example same IPC id may refer to different objects
1176 or same user id or pid may refer to different tasks when used in
1177 different namespaces.
1182 bool "UTS namespace"
1185 In this namespace tasks see different info provided with the
1189 bool "IPC namespace"
1190 depends on (SYSVIPC || POSIX_MQUEUE)
1193 In this namespace tasks work with IPC ids which correspond to
1194 different IPC objects in different namespaces.
1197 bool "User namespace"
1200 This allows containers, i.e. vservers, to use user namespaces
1201 to provide different user info for different servers.
1203 When user namespaces are enabled in the kernel it is
1204 recommended that the MEMCG option also be enabled and that
1205 user-space use the memory control groups to limit the amount
1206 of memory a memory unprivileged users can use.
1211 bool "PID Namespaces"
1214 Support process id namespaces. This allows having multiple
1215 processes with the same pid as long as they are in different
1216 pid namespaces. This is a building block of containers.
1219 bool "Network namespace"
1223 Allow user space to create what appear to be multiple instances
1224 of the network stack.
1228 config SCHED_AUTOGROUP
1229 bool "Automatic process group scheduling"
1232 select FAIR_GROUP_SCHED
1234 This option optimizes the scheduler for common desktop workloads by
1235 automatically creating and populating task groups. This separation
1236 of workloads isolates aggressive CPU burners (like build jobs) from
1237 desktop applications. Task group autogeneration is currently based
1240 config SYSFS_DEPRECATED
1241 bool "Enable deprecated sysfs features to support old userspace tools"
1245 This option adds code that switches the layout of the "block" class
1246 devices, to not show up in /sys/class/block/, but only in
1249 This switch is only active when the sysfs.deprecated=1 boot option is
1250 passed or the SYSFS_DEPRECATED_V2 option is set.
1252 This option allows new kernels to run on old distributions and tools,
1253 which might get confused by /sys/class/block/. Since 2007/2008 all
1254 major distributions and tools handle this just fine.
1256 Recent distributions and userspace tools after 2009/2010 depend on
1257 the existence of /sys/class/block/, and will not work with this
1260 Only if you are using a new kernel on an old distribution, you might
1263 config SYSFS_DEPRECATED_V2
1264 bool "Enable deprecated sysfs features by default"
1267 depends on SYSFS_DEPRECATED
1269 Enable deprecated sysfs by default.
1271 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1274 Only if you are using a new kernel on an old distribution, you might
1275 need to say Y here. Even then, odds are you would not need it
1276 enabled, you can always pass the boot option if absolutely necessary.
1279 bool "Kernel->user space relay support (formerly relayfs)"
1281 This option enables support for relay interface support in
1282 certain file systems (such as debugfs).
1283 It is designed to provide an efficient mechanism for tools and
1284 facilities to relay large amounts of data from kernel space to
1289 config BLK_DEV_INITRD
1290 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1291 depends on BROKEN || !FRV
1293 The initial RAM filesystem is a ramfs which is loaded by the
1294 boot loader (loadlin or lilo) and that is mounted as root
1295 before the normal boot procedure. It is typically used to
1296 load modules needed to mount the "real" root file system,
1297 etc. See <file:Documentation/initrd.txt> for details.
1299 If RAM disk support (BLK_DEV_RAM) is also included, this
1300 also enables initial RAM disk (initrd) support and adds
1301 15 Kbytes (more on some other architectures) to the kernel size.
1307 source "usr/Kconfig"
1312 prompt "Compiler optimization level"
1313 default CONFIG_CC_OPTIMIZE_FOR_PERFORMANCE
1315 config CC_OPTIMIZE_FOR_PERFORMANCE
1316 bool "Optimize for performance"
1318 This is the default optimization level for the kernel, building
1319 with the "-O2" compiler flag for best performance and most
1320 helpful compile-time warnings.
1322 config CC_OPTIMIZE_FOR_SIZE
1323 bool "Optimize for size"
1325 Enabling this option will pass "-Os" instead of "-O2" to
1326 your compiler resulting in a smaller kernel.
1341 config SYSCTL_EXCEPTION_TRACE
1344 Enable support for /proc/sys/debug/exception-trace.
1346 config SYSCTL_ARCH_UNALIGN_NO_WARN
1349 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1350 Allows arch to define/use @no_unaligned_warning to possibly warn
1351 about unaligned access emulation going on under the hood.
1353 config SYSCTL_ARCH_UNALIGN_ALLOW
1356 Enable support for /proc/sys/kernel/unaligned-trap
1357 Allows arches to define/use @unaligned_enabled to runtime toggle
1358 the unaligned access emulation.
1359 see arch/parisc/kernel/unaligned.c for reference
1361 config HAVE_PCSPKR_PLATFORM
1364 # interpreter that classic socket filters depend on
1369 bool "Configure standard kernel features (expert users)"
1370 # Unhide debug options, to make the on-by-default options visible
1373 This option allows certain base kernel options and settings
1374 to be disabled or tweaked. This is for specialized
1375 environments which can tolerate a "non-standard" kernel.
1376 Only use this if you really know what you are doing.
1379 bool "Enable 16-bit UID system calls" if EXPERT
1380 depends on HAVE_UID16 && MULTIUSER
1383 This enables the legacy 16-bit UID syscall wrappers.
1386 bool "Multiple users, groups and capabilities support" if EXPERT
1389 This option enables support for non-root users, groups and
1392 If you say N here, all processes will run with UID 0, GID 0, and all
1393 possible capabilities. Saying N here also compiles out support for
1394 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1397 If unsure, say Y here.
1399 config SGETMASK_SYSCALL
1400 bool "sgetmask/ssetmask syscalls support" if EXPERT
1401 def_bool PARISC || MN10300 || BLACKFIN || M68K || PPC || MIPS || X86 || SPARC || CRIS || MICROBLAZE || SUPERH
1403 sys_sgetmask and sys_ssetmask are obsolete system calls
1404 no longer supported in libc but still enabled by default in some
1407 If unsure, leave the default option here.
1409 config SYSFS_SYSCALL
1410 bool "Sysfs syscall support" if EXPERT
1413 sys_sysfs is an obsolete system call no longer supported in libc.
1414 Note that disabling this option is more secure but might break
1415 compatibility with some systems.
1417 If unsure say Y here.
1419 config SYSCTL_SYSCALL
1420 bool "Sysctl syscall support" if EXPERT
1421 depends on PROC_SYSCTL
1425 sys_sysctl uses binary paths that have been found challenging
1426 to properly maintain and use. The interface in /proc/sys
1427 using paths with ascii names is now the primary path to this
1430 Almost nothing using the binary sysctl interface so if you are
1431 trying to save some space it is probably safe to disable this,
1432 making your kernel marginally smaller.
1434 If unsure say N here.
1437 bool "Load all symbols for debugging/ksymoops" if EXPERT
1440 Say Y here to let the kernel print out symbolic crash information and
1441 symbolic stack backtraces. This increases the size of the kernel
1442 somewhat, as all symbols have to be loaded into the kernel image.
1445 bool "Include all symbols in kallsyms"
1446 depends on DEBUG_KERNEL && KALLSYMS
1448 Normally kallsyms only contains the symbols of functions for nicer
1449 OOPS messages and backtraces (i.e., symbols from the text and inittext
1450 sections). This is sufficient for most cases. And only in very rare
1451 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1452 names of variables from the data sections, etc).
1454 This option makes sure that all symbols are loaded into the kernel
1455 image (i.e., symbols from all sections) in cost of increased kernel
1456 size (depending on the kernel configuration, it may be 300KiB or
1457 something like this).
1459 Say N unless you really need all symbols.
1461 config KALLSYMS_ABSOLUTE_PERCPU
1464 default X86_64 && SMP
1466 config KALLSYMS_BASE_RELATIVE
1469 default !IA64 && !(TILE && 64BIT)
1471 Instead of emitting them as absolute values in the native word size,
1472 emit the symbol references in the kallsyms table as 32-bit entries,
1473 each containing a relative value in the range [base, base + U32_MAX]
1474 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1475 an absolute value in the range [0, S32_MAX] or a relative value in the
1476 range [base, base + S32_MAX], where base is the lowest relative symbol
1477 address encountered in the image.
1479 On 64-bit builds, this reduces the size of the address table by 50%,
1480 but more importantly, it results in entries whose values are build
1481 time constants, and no relocation pass is required at runtime to fix
1482 up the entries based on the runtime load address of the kernel.
1486 bool "Enable support for printk" if EXPERT
1489 This option enables normal printk support. Removing it
1490 eliminates most of the message strings from the kernel image
1491 and makes the kernel more or less silent. As this makes it
1492 very difficult to diagnose system problems, saying N here is
1493 strongly discouraged.
1501 bool "BUG() support" if EXPERT
1504 Disabling this option eliminates support for BUG and WARN, reducing
1505 the size of your kernel image and potentially quietly ignoring
1506 numerous fatal conditions. You should only consider disabling this
1507 option for embedded systems with no facilities for reporting errors.
1513 bool "Enable ELF core dumps" if EXPERT
1515 Enable support for generating core dumps. Disabling saves about 4k.
1518 config PCSPKR_PLATFORM
1519 bool "Enable PC-Speaker support" if EXPERT
1520 depends on HAVE_PCSPKR_PLATFORM
1524 This option allows to disable the internal PC-Speaker
1525 support, saving some memory.
1529 bool "Enable full-sized data structures for core" if EXPERT
1531 Disabling this option reduces the size of miscellaneous core
1532 kernel data structures. This saves memory on small machines,
1533 but may reduce performance.
1536 bool "Enable futex support" if EXPERT
1540 Disabling this option will cause the kernel to be built without
1541 support for "fast userspace mutexes". The resulting kernel may not
1542 run glibc-based applications correctly.
1544 config HAVE_FUTEX_CMPXCHG
1548 Architectures should select this if futex_atomic_cmpxchg_inatomic()
1549 is implemented and always working. This removes a couple of runtime
1553 bool "Enable eventpoll support" if EXPERT
1557 Disabling this option will cause the kernel to be built without
1558 support for epoll family of system calls.
1561 bool "Enable signalfd() system call" if EXPERT
1565 Enable the signalfd() system call that allows to receive signals
1566 on a file descriptor.
1571 bool "Enable timerfd() system call" if EXPERT
1575 Enable the timerfd() system call that allows to receive timer
1576 events on a file descriptor.
1581 bool "Enable eventfd() system call" if EXPERT
1585 Enable the eventfd() system call that allows to receive both
1586 kernel notification (ie. KAIO) or userspace notifications.
1590 # syscall, maps, verifier
1592 bool "Enable bpf() system call"
1597 Enable the bpf() system call that allows to manipulate eBPF
1598 programs and maps via file descriptors.
1601 bool "Use full shmem filesystem" if EXPERT
1605 The shmem is an internal filesystem used to manage shared memory.
1606 It is backed by swap and manages resource limits. It is also exported
1607 to userspace as tmpfs if TMPFS is enabled. Disabling this
1608 option replaces shmem and tmpfs with the much simpler ramfs code,
1609 which may be appropriate on small systems without swap.
1612 bool "Enable AIO support" if EXPERT
1615 This option enables POSIX asynchronous I/O which may by used
1616 by some high performance threaded applications. Disabling
1617 this option saves about 7k.
1619 config ADVISE_SYSCALLS
1620 bool "Enable madvise/fadvise syscalls" if EXPERT
1623 This option enables the madvise and fadvise syscalls, used by
1624 applications to advise the kernel about their future memory or file
1625 usage, improving performance. If building an embedded system where no
1626 applications use these syscalls, you can disable this option to save
1630 bool "Enable userfaultfd() system call"
1634 Enable the userfaultfd() system call that allows to intercept and
1635 handle page faults in userland.
1639 bool "Enable PCI quirk workarounds" if EXPERT
1642 This enables workarounds for various PCI chipset
1643 bugs/quirks. Disable this only if your target machine is
1644 unaffected by PCI quirks.
1647 bool "Enable membarrier() system call" if EXPERT
1650 Enable the membarrier() system call that allows issuing memory
1651 barriers across all running threads, which can be used to distribute
1652 the cost of user-space memory barriers asymmetrically by transforming
1653 pairs of memory barriers into pairs consisting of membarrier() and a
1659 bool "Embedded system"
1660 option allnoconfig_y
1663 This option should be enabled if compiling the kernel for
1664 an embedded system so certain expert options are available
1667 config HAVE_PERF_EVENTS
1670 See tools/perf/design.txt for details.
1672 config PERF_USE_VMALLOC
1675 See tools/perf/design.txt for details
1677 menu "Kernel Performance Events And Counters"
1680 bool "Kernel performance events and counters"
1681 default y if PROFILING
1682 depends on HAVE_PERF_EVENTS
1687 Enable kernel support for various performance events provided
1688 by software and hardware.
1690 Software events are supported either built-in or via the
1691 use of generic tracepoints.
1693 Most modern CPUs support performance events via performance
1694 counter registers. These registers count the number of certain
1695 types of hw events: such as instructions executed, cachemisses
1696 suffered, or branches mis-predicted - without slowing down the
1697 kernel or applications. These registers can also trigger interrupts
1698 when a threshold number of events have passed - and can thus be
1699 used to profile the code that runs on that CPU.
1701 The Linux Performance Event subsystem provides an abstraction of
1702 these software and hardware event capabilities, available via a
1703 system call and used by the "perf" utility in tools/perf/. It
1704 provides per task and per CPU counters, and it provides event
1705 capabilities on top of those.
1709 config DEBUG_PERF_USE_VMALLOC
1711 bool "Debug: use vmalloc to back perf mmap() buffers"
1712 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1713 select PERF_USE_VMALLOC
1715 Use vmalloc memory to back perf mmap() buffers.
1717 Mostly useful for debugging the vmalloc code on platforms
1718 that don't require it.
1724 config VM_EVENT_COUNTERS
1726 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1728 VM event counters are needed for event counts to be shown.
1729 This option allows the disabling of the VM event counters
1730 on EXPERT systems. /proc/vmstat will only show page counts
1731 if VM event counters are disabled.
1735 bool "Enable SLUB debugging support" if EXPERT
1736 depends on SLUB && SYSFS
1738 SLUB has extensive debug support features. Disabling these can
1739 result in significant savings in code size. This also disables
1740 SLUB sysfs support. /sys/slab will not exist and there will be
1741 no support for cache validation etc.
1744 bool "Disable heap randomization"
1747 Randomizing heap placement makes heap exploits harder, but it
1748 also breaks ancient binaries (including anything libc5 based).
1749 This option changes the bootup default to heap randomization
1750 disabled, and can be overridden at runtime by setting
1751 /proc/sys/kernel/randomize_va_space to 2.
1753 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1756 prompt "Choose SLAB allocator"
1759 This option allows to select a slab allocator.
1764 The regular slab allocator that is established and known to work
1765 well in all environments. It organizes cache hot objects in
1766 per cpu and per node queues.
1769 bool "SLUB (Unqueued Allocator)"
1771 SLUB is a slab allocator that minimizes cache line usage
1772 instead of managing queues of cached objects (SLAB approach).
1773 Per cpu caching is realized using slabs of objects instead
1774 of queues of objects. SLUB can use memory efficiently
1775 and has enhanced diagnostics. SLUB is the default choice for
1780 bool "SLOB (Simple Allocator)"
1782 SLOB replaces the stock allocator with a drastically simpler
1783 allocator. SLOB is generally more space efficient but
1784 does not perform as well on large systems.
1788 config SLAB_FREELIST_RANDOM
1790 depends on SLAB || SLUB
1791 bool "SLAB freelist randomization"
1793 Randomizes the freelist order used on creating new pages. This
1794 security feature reduces the predictability of the kernel slab
1795 allocator against heap overflows.
1797 config SLUB_CPU_PARTIAL
1799 depends on SLUB && SMP
1800 bool "SLUB per cpu partial cache"
1802 Per cpu partial caches accellerate objects allocation and freeing
1803 that is local to a processor at the price of more indeterminism
1804 in the latency of the free. On overflow these caches will be cleared
1805 which requires the taking of locks that may cause latency spikes.
1806 Typically one would choose no for a realtime system.
1808 config MMAP_ALLOW_UNINITIALIZED
1809 bool "Allow mmapped anonymous memory to be uninitialized"
1810 depends on EXPERT && !MMU
1813 Normally, and according to the Linux spec, anonymous memory obtained
1814 from mmap() has it's contents cleared before it is passed to
1815 userspace. Enabling this config option allows you to request that
1816 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1817 providing a huge performance boost. If this option is not enabled,
1818 then the flag will be ignored.
1820 This is taken advantage of by uClibc's malloc(), and also by
1821 ELF-FDPIC binfmt's brk and stack allocator.
1823 Because of the obvious security issues, this option should only be
1824 enabled on embedded devices where you control what is run in
1825 userspace. Since that isn't generally a problem on no-MMU systems,
1826 it is normally safe to say Y here.
1828 See Documentation/nommu-mmap.txt for more information.
1830 config SYSTEM_DATA_VERIFICATION
1832 select SYSTEM_TRUSTED_KEYRING
1836 select ASYMMETRIC_KEY_TYPE
1837 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1840 select X509_CERTIFICATE_PARSER
1841 select PKCS7_MESSAGE_PARSER
1843 Provide PKCS#7 message verification using the contents of the system
1844 trusted keyring to provide public keys. This then can be used for
1845 module verification, kexec image verification and firmware blob
1849 bool "Profiling support"
1851 Say Y here to enable the extended profiling support mechanisms used
1852 by profilers such as OProfile.
1855 # Place an empty function call at each tracepoint site. Can be
1856 # dynamically changed for a probe function.
1861 source "arch/Kconfig"
1863 endmenu # General setup
1865 config HAVE_GENERIC_DMA_COHERENT
1872 depends on SLAB || SLUB_DEBUG
1880 default 0 if BASE_FULL
1881 default 1 if !BASE_FULL
1884 bool "Enable loadable module support"
1887 Kernel modules are small pieces of compiled code which can
1888 be inserted in the running kernel, rather than being
1889 permanently built into the kernel. You use the "modprobe"
1890 tool to add (and sometimes remove) them. If you say Y here,
1891 many parts of the kernel can be built as modules (by
1892 answering M instead of Y where indicated): this is most
1893 useful for infrequently used options which are not required
1894 for booting. For more information, see the man pages for
1895 modprobe, lsmod, modinfo, insmod and rmmod.
1897 If you say Y here, you will need to run "make
1898 modules_install" to put the modules under /lib/modules/
1899 where modprobe can find them (you may need to be root to do
1906 config MODULE_FORCE_LOAD
1907 bool "Forced module loading"
1910 Allow loading of modules without version information (ie. modprobe
1911 --force). Forced module loading sets the 'F' (forced) taint flag and
1912 is usually a really bad idea.
1914 config MODULE_UNLOAD
1915 bool "Module unloading"
1917 Without this option you will not be able to unload any
1918 modules (note that some modules may not be unloadable
1919 anyway), which makes your kernel smaller, faster
1920 and simpler. If unsure, say Y.
1922 config MODULE_FORCE_UNLOAD
1923 bool "Forced module unloading"
1924 depends on MODULE_UNLOAD
1926 This option allows you to force a module to unload, even if the
1927 kernel believes it is unsafe: the kernel will remove the module
1928 without waiting for anyone to stop using it (using the -f option to
1929 rmmod). This is mainly for kernel developers and desperate users.
1933 bool "Module versioning support"
1935 Usually, you have to use modules compiled with your kernel.
1936 Saying Y here makes it sometimes possible to use modules
1937 compiled for different kernels, by adding enough information
1938 to the modules to (hopefully) spot any changes which would
1939 make them incompatible with the kernel you are running. If
1942 config MODULE_SRCVERSION_ALL
1943 bool "Source checksum for all modules"
1945 Modules which contain a MODULE_VERSION get an extra "srcversion"
1946 field inserted into their modinfo section, which contains a
1947 sum of the source files which made it. This helps maintainers
1948 see exactly which source was used to build a module (since
1949 others sometimes change the module source without updating
1950 the version). With this option, such a "srcversion" field
1951 will be created for all modules. If unsure, say N.
1954 bool "Module signature verification"
1956 select SYSTEM_DATA_VERIFICATION
1958 Check modules for valid signatures upon load: the signature
1959 is simply appended to the module. For more information see
1960 Documentation/module-signing.txt.
1962 Note that this option adds the OpenSSL development packages as a
1963 kernel build dependency so that the signing tool can use its crypto
1966 !!!WARNING!!! If you enable this option, you MUST make sure that the
1967 module DOES NOT get stripped after being signed. This includes the
1968 debuginfo strip done by some packagers (such as rpmbuild) and
1969 inclusion into an initramfs that wants the module size reduced.
1971 config MODULE_SIG_FORCE
1972 bool "Require modules to be validly signed"
1973 depends on MODULE_SIG
1975 Reject unsigned modules or signed modules for which we don't have a
1976 key. Without this, such modules will simply taint the kernel.
1978 config MODULE_SIG_ALL
1979 bool "Automatically sign all modules"
1981 depends on MODULE_SIG
1983 Sign all modules during make modules_install. Without this option,
1984 modules must be signed manually, using the scripts/sign-file tool.
1986 comment "Do not forget to sign required modules with scripts/sign-file"
1987 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
1990 prompt "Which hash algorithm should modules be signed with?"
1991 depends on MODULE_SIG
1993 This determines which sort of hashing algorithm will be used during
1994 signature generation. This algorithm _must_ be built into the kernel
1995 directly so that signature verification can take place. It is not
1996 possible to load a signed module containing the algorithm to check
1997 the signature on that module.
1999 config MODULE_SIG_SHA1
2000 bool "Sign modules with SHA-1"
2003 config MODULE_SIG_SHA224
2004 bool "Sign modules with SHA-224"
2005 select CRYPTO_SHA256
2007 config MODULE_SIG_SHA256
2008 bool "Sign modules with SHA-256"
2009 select CRYPTO_SHA256
2011 config MODULE_SIG_SHA384
2012 bool "Sign modules with SHA-384"
2013 select CRYPTO_SHA512
2015 config MODULE_SIG_SHA512
2016 bool "Sign modules with SHA-512"
2017 select CRYPTO_SHA512
2021 config MODULE_SIG_HASH
2023 depends on MODULE_SIG
2024 default "sha1" if MODULE_SIG_SHA1
2025 default "sha224" if MODULE_SIG_SHA224
2026 default "sha256" if MODULE_SIG_SHA256
2027 default "sha384" if MODULE_SIG_SHA384
2028 default "sha512" if MODULE_SIG_SHA512
2030 config MODULE_COMPRESS
2031 bool "Compress modules on installation"
2035 Compresses kernel modules when 'make modules_install' is run; gzip or
2036 xz depending on "Compression algorithm" below.
2038 module-init-tools MAY support gzip, and kmod MAY support gzip and xz.
2040 Out-of-tree kernel modules installed using Kbuild will also be
2041 compressed upon installation.
2043 Note: for modules inside an initrd or initramfs, it's more efficient
2044 to compress the whole initrd or initramfs instead.
2046 Note: This is fully compatible with signed modules.
2051 prompt "Compression algorithm"
2052 depends on MODULE_COMPRESS
2053 default MODULE_COMPRESS_GZIP
2055 This determines which sort of compression will be used during
2056 'make modules_install'.
2058 GZIP (default) and XZ are supported.
2060 config MODULE_COMPRESS_GZIP
2063 config MODULE_COMPRESS_XZ
2068 config TRIM_UNUSED_KSYMS
2069 bool "Trim unused exported kernel symbols"
2070 depends on MODULES && !UNUSED_SYMBOLS
2072 The kernel and some modules make many symbols available for
2073 other modules to use via EXPORT_SYMBOL() and variants. Depending
2074 on the set of modules being selected in your kernel configuration,
2075 many of those exported symbols might never be used.
2077 This option allows for unused exported symbols to be dropped from
2078 the build. In turn, this provides the compiler more opportunities
2079 (especially when using LTO) for optimizing the code and reducing
2080 binary size. This might have some security advantages as well.
2086 config MODULES_TREE_LOOKUP
2088 depends on PERF_EVENTS || TRACING
2090 config INIT_ALL_POSSIBLE
2093 Back when each arch used to define their own cpu_online_mask and
2094 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2095 with all 1s, and others with all 0s. When they were centralised,
2096 it was better to provide this option than to break all the archs
2097 and have several arch maintainers pursuing me down dark alleys.
2099 source "block/Kconfig"
2101 config PREEMPT_NOTIFIERS
2108 # Can be selected by architectures with broken toolchains
2109 # that get confused by correct const<->read_only section
2111 config BROKEN_RODATA
2117 Build a simple ASN.1 grammar compiler that produces a bytecode output
2118 that can be interpreted by the ASN.1 stream decoder and used to
2119 inform it as to what tags are to be expected in a stream and what
2120 functions to call on what tags.
2122 source "kernel/Kconfig.locks"