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
32 bool "Prompt for development and/or incomplete code/drivers"
34 Some of the various things that Linux supports (such as network
35 drivers, file systems, network protocols, etc.) can be in a state
36 of development where the functionality, stability, or the level of
37 testing is not yet high enough for general use. This is usually
38 known as the "alpha-test" phase among developers. If a feature is
39 currently in alpha-test, then the developers usually discourage
40 uninformed widespread use of this feature by the general public to
41 avoid "Why doesn't this work?" type mail messages. However, active
42 testing and use of these systems is welcomed. Just be aware that it
43 may not meet the normal level of reliability or it may fail to work
44 in some special cases. Detailed bug reports from people familiar
45 with the kernel internals are usually welcomed by the developers
46 (before submitting bug reports, please read the documents
47 <file:README>, <file:MAINTAINERS>, <file:REPORTING-BUGS>,
48 <file:Documentation/BUG-HUNTING>, and
49 <file:Documentation/oops-tracing.txt> in the kernel source).
51 This option will also make obsoleted drivers available. These are
52 drivers that have been replaced by something else, and/or are
53 scheduled to be removed in a future kernel release.
55 Unless you intend to help test and develop a feature or driver that
56 falls into this category, or you have a situation that requires
57 using these features, you should probably say N here, which will
58 cause the configurator to present you with fewer choices. If
59 you say Y here, you will be offered the choice of using features or
60 drivers that are currently considered to be in the alpha-test phase.
67 depends on BROKEN || !SMP
70 config INIT_ENV_ARG_LIMIT
75 Maximum of each of the number of arguments and environment
76 variables passed to init from the kernel command line.
80 string "Cross-compiler tool prefix"
82 Same as running 'make CROSS_COMPILE=prefix-' but stored for
83 default make runs in this kernel build directory. You don't
84 need to set this unless you want the configured kernel build
85 directory to select the cross-compiler automatically.
88 string "Local version - append to kernel release"
90 Append an extra string to the end of your kernel version.
91 This will show up when you type uname, for example.
92 The string you set here will be appended after the contents of
93 any files with a filename matching localversion* in your
94 object and source tree, in that order. Your total string can
95 be a maximum of 64 characters.
97 config LOCALVERSION_AUTO
98 bool "Automatically append version information to the version string"
101 This will try to automatically determine if the current tree is a
102 release tree by looking for git tags that belong to the current
103 top of tree revision.
105 A string of the format -gxxxxxxxx will be added to the localversion
106 if a git-based tree is found. The string generated by this will be
107 appended after any matching localversion* files, and after the value
108 set in CONFIG_LOCALVERSION.
110 (The actual string used here is the first eight characters produced
111 by running the command:
113 $ git rev-parse --verify HEAD
115 which is done within the script "scripts/setlocalversion".)
117 config HAVE_KERNEL_GZIP
120 config HAVE_KERNEL_BZIP2
123 config HAVE_KERNEL_LZMA
126 config HAVE_KERNEL_XZ
129 config HAVE_KERNEL_LZO
133 prompt "Kernel compression mode"
135 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO
137 The linux kernel is a kind of self-extracting executable.
138 Several compression algorithms are available, which differ
139 in efficiency, compression and decompression speed.
140 Compression speed is only relevant when building a kernel.
141 Decompression speed is relevant at each boot.
143 If you have any problems with bzip2 or lzma compressed
144 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
145 version of this functionality (bzip2 only), for 2.4, was
146 supplied by Christian Ludwig)
148 High compression options are mostly useful for users, who
149 are low on disk space (embedded systems), but for whom ram
152 If in doubt, select 'gzip'
156 depends on HAVE_KERNEL_GZIP
158 The old and tried gzip compression. It provides a good balance
159 between compression ratio and decompression speed.
163 depends on HAVE_KERNEL_BZIP2
165 Its compression ratio and speed is intermediate.
166 Decompression speed is slowest among the choices. The kernel
167 size is about 10% smaller with bzip2, in comparison to gzip.
168 Bzip2 uses a large amount of memory. For modern kernels you
169 will need at least 8MB RAM or more for booting.
173 depends on HAVE_KERNEL_LZMA
175 This compression algorithm's ratio is best. Decompression speed
176 is between gzip and bzip2. Compression is slowest.
177 The kernel size is about 33% smaller with LZMA in comparison to gzip.
181 depends on HAVE_KERNEL_XZ
183 XZ uses the LZMA2 algorithm and instruction set specific
184 BCJ filters which can improve compression ratio of executable
185 code. The size of the kernel is about 30% smaller with XZ in
186 comparison to gzip. On architectures for which there is a BCJ
187 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
188 will create a few percent smaller kernel than plain LZMA.
190 The speed is about the same as with LZMA: The decompression
191 speed of XZ is better than that of bzip2 but worse than gzip
192 and LZO. Compression is slow.
196 depends on HAVE_KERNEL_LZO
198 Its compression ratio is the poorest among the choices. The kernel
199 size is about 10% bigger than gzip; however its speed
200 (both compression and decompression) is the fastest.
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"
246 depends on NET && EXPERIMENTAL
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
267 bool "open by fhandle syscalls"
270 If you say Y here, a user level program will be able to map
271 file names to handle and then later use the handle for
272 different file system operations. This is useful in implementing
273 userspace file servers, which now track files using handles instead
274 of names. The handle would remain the same even if file names
275 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
279 bool "Auditing support"
282 Enable auditing infrastructure that can be used with another
283 kernel subsystem, such as SELinux (which requires this for
284 logging of avc messages output). Does not do system-call
285 auditing without CONFIG_AUDITSYSCALL.
288 bool "Enable system-call auditing support"
289 depends on AUDIT && (X86 || PPC || S390 || IA64 || UML || SPARC64 || SUPERH || (ARM && AEABI && !OABI_COMPAT))
290 default y if SECURITY_SELINUX
292 Enable low-overhead system-call auditing infrastructure that
293 can be used independently or with another kernel subsystem,
298 depends on AUDITSYSCALL
303 depends on AUDITSYSCALL
306 config AUDIT_LOGINUID_IMMUTABLE
307 bool "Make audit loginuid immutable"
310 The config option toggles if a task setting its loginuid requires
311 CAP_SYS_AUDITCONTROL or if that task should require no special permissions
312 but should instead only allow setting its loginuid if it was never
313 previously set. On systems which use systemd or a similar central
314 process to restart login services this should be set to true. On older
315 systems in which an admin would typically have to directly stop and
316 start processes this should be set to false. Setting this to true allows
317 one to drop potentially dangerous capabilites from the login tasks,
318 but may not be backwards compatible with older init systems.
320 source "kernel/irq/Kconfig"
321 source "kernel/time/Kconfig"
323 menu "CPU/Task time and stats accounting"
326 prompt "Cputime accounting"
327 default TICK_CPU_ACCOUNTING if !PPC64
328 default VIRT_CPU_ACCOUNTING if PPC64
330 # Kind of a stub config for the pure tick based cputime accounting
331 config TICK_CPU_ACCOUNTING
332 bool "Simple tick based cputime accounting"
335 This is the basic tick based cputime accounting that maintains
336 statistics about user, system and idle time spent on per jiffies
341 config VIRT_CPU_ACCOUNTING
342 bool "Deterministic task and CPU time accounting"
343 depends on HAVE_VIRT_CPU_ACCOUNTING
345 Select this option to enable more accurate task and CPU time
346 accounting. This is done by reading a CPU counter on each
347 kernel entry and exit and on transitions within the kernel
348 between system, softirq and hardirq state, so there is a
349 small performance impact. In the case of s390 or IBM POWER > 5,
350 this also enables accounting of stolen time on logically-partitioned
353 config IRQ_TIME_ACCOUNTING
354 bool "Fine granularity task level IRQ time accounting"
355 depends on HAVE_IRQ_TIME_ACCOUNTING
357 Select this option to enable fine granularity task irq time
358 accounting. This is done by reading a timestamp on each
359 transitions between softirq and hardirq state, so there can be a
360 small performance impact.
362 If in doubt, say N here.
366 config BSD_PROCESS_ACCT
367 bool "BSD Process Accounting"
369 If you say Y here, a user level program will be able to instruct the
370 kernel (via a special system call) to write process accounting
371 information to a file: whenever a process exits, information about
372 that process will be appended to the file by the kernel. The
373 information includes things such as creation time, owning user,
374 command name, memory usage, controlling terminal etc. (the complete
375 list is in the struct acct in <file:include/linux/acct.h>). It is
376 up to the user level program to do useful things with this
377 information. This is generally a good idea, so say Y.
379 config BSD_PROCESS_ACCT_V3
380 bool "BSD Process Accounting version 3 file format"
381 depends on BSD_PROCESS_ACCT
384 If you say Y here, the process accounting information is written
385 in a new file format that also logs the process IDs of each
386 process and it's parent. Note that this file format is incompatible
387 with previous v0/v1/v2 file formats, so you will need updated tools
388 for processing it. A preliminary version of these tools is available
389 at <http://www.gnu.org/software/acct/>.
392 bool "Export task/process statistics through netlink (EXPERIMENTAL)"
396 Export selected statistics for tasks/processes through the
397 generic netlink interface. Unlike BSD process accounting, the
398 statistics are available during the lifetime of tasks/processes as
399 responses to commands. Like BSD accounting, they are sent to user
404 config TASK_DELAY_ACCT
405 bool "Enable per-task delay accounting (EXPERIMENTAL)"
408 Collect information on time spent by a task waiting for system
409 resources like cpu, synchronous block I/O completion and swapping
410 in pages. Such statistics can help in setting a task's priorities
411 relative to other tasks for cpu, io, rss limits etc.
416 bool "Enable extended accounting over taskstats (EXPERIMENTAL)"
419 Collect extended task accounting data and send the data
420 to userland for processing over the taskstats interface.
424 config TASK_IO_ACCOUNTING
425 bool "Enable per-task storage I/O accounting (EXPERIMENTAL)"
426 depends on TASK_XACCT
428 Collect information on the number of bytes of storage I/O which this
433 endmenu # "CPU/Task time and stats accounting"
438 prompt "RCU Implementation"
442 bool "Tree-based hierarchical RCU"
443 depends on !PREEMPT && SMP
445 This option selects the RCU implementation that is
446 designed for very large SMP system with hundreds or
447 thousands of CPUs. It also scales down nicely to
450 config TREE_PREEMPT_RCU
451 bool "Preemptible tree-based hierarchical RCU"
452 depends on PREEMPT && SMP
454 This option selects the RCU implementation that is
455 designed for very large SMP systems with hundreds or
456 thousands of CPUs, but for which real-time response
457 is also required. It also scales down nicely to
461 bool "UP-only small-memory-footprint RCU"
462 depends on !PREEMPT && !SMP
464 This option selects the RCU implementation that is
465 designed for UP systems from which real-time response
466 is not required. This option greatly reduces the
467 memory footprint of RCU.
469 config TINY_PREEMPT_RCU
470 bool "Preemptible UP-only small-memory-footprint RCU"
471 depends on PREEMPT && !SMP
473 This option selects the RCU implementation that is designed
474 for real-time UP systems. This option greatly reduces the
475 memory footprint of RCU.
480 def_bool ( TREE_PREEMPT_RCU || TINY_PREEMPT_RCU )
482 This option enables preemptible-RCU code that is common between
483 the TREE_PREEMPT_RCU and TINY_PREEMPT_RCU implementations.
485 config CONTEXT_TRACKING
489 bool "Consider userspace as in RCU extended quiescent state"
490 depends on HAVE_CONTEXT_TRACKING && SMP
491 select CONTEXT_TRACKING
493 This option sets hooks on kernel / userspace boundaries and
494 puts RCU in extended quiescent state when the CPU runs in
495 userspace. It means that when a CPU runs in userspace, it is
496 excluded from the global RCU state machine and thus doesn't
497 try to keep the timer tick on for RCU.
499 Unless you want to hack and help the development of the full
500 dynticks mode, you shouldn't enable this option. It also
501 adds unnecessary overhead.
505 config CONTEXT_TRACKING_FORCE
506 bool "Force context tracking"
507 depends on CONTEXT_TRACKING
509 Probe on user/kernel boundaries by default in order to
510 test the features that rely on it such as userspace RCU extended
512 This test is there for debugging until we have a real user like the
516 int "Tree-based hierarchical RCU fanout value"
519 depends on TREE_RCU || TREE_PREEMPT_RCU
523 This option controls the fanout of hierarchical implementations
524 of RCU, allowing RCU to work efficiently on machines with
525 large numbers of CPUs. This value must be at least the fourth
526 root of NR_CPUS, which allows NR_CPUS to be insanely large.
527 The default value of RCU_FANOUT should be used for production
528 systems, but if you are stress-testing the RCU implementation
529 itself, small RCU_FANOUT values allow you to test large-system
530 code paths on small(er) systems.
532 Select a specific number if testing RCU itself.
533 Take the default if unsure.
535 config RCU_FANOUT_LEAF
536 int "Tree-based hierarchical RCU leaf-level fanout value"
537 range 2 RCU_FANOUT if 64BIT
538 range 2 RCU_FANOUT if !64BIT
539 depends on TREE_RCU || TREE_PREEMPT_RCU
542 This option controls the leaf-level fanout of hierarchical
543 implementations of RCU, and allows trading off cache misses
544 against lock contention. Systems that synchronize their
545 scheduling-clock interrupts for energy-efficiency reasons will
546 want the default because the smaller leaf-level fanout keeps
547 lock contention levels acceptably low. Very large systems
548 (hundreds or thousands of CPUs) will instead want to set this
549 value to the maximum value possible in order to reduce the
550 number of cache misses incurred during RCU's grace-period
551 initialization. These systems tend to run CPU-bound, and thus
552 are not helped by synchronized interrupts, and thus tend to
553 skew them, which reduces lock contention enough that large
554 leaf-level fanouts work well.
556 Select a specific number if testing RCU itself.
558 Select the maximum permissible value for large systems.
560 Take the default if unsure.
562 config RCU_FANOUT_EXACT
563 bool "Disable tree-based hierarchical RCU auto-balancing"
564 depends on TREE_RCU || TREE_PREEMPT_RCU
567 This option forces use of the exact RCU_FANOUT value specified,
568 regardless of imbalances in the hierarchy. This is useful for
569 testing RCU itself, and might one day be useful on systems with
570 strong NUMA behavior.
572 Without RCU_FANOUT_EXACT, the code will balance the hierarchy.
576 config RCU_FAST_NO_HZ
577 bool "Accelerate last non-dyntick-idle CPU's grace periods"
578 depends on NO_HZ && SMP
581 This option causes RCU to attempt to accelerate grace periods in
582 order to allow CPUs to enter dynticks-idle state more quickly.
583 On the other hand, this option increases the overhead of the
584 dynticks-idle checking, thus degrading scheduling latency.
586 Say Y if energy efficiency is critically important, and you don't
587 care about real-time response.
589 Say N if you are unsure.
591 config TREE_RCU_TRACE
592 def_bool RCU_TRACE && ( TREE_RCU || TREE_PREEMPT_RCU )
595 This option provides tracing for the TREE_RCU and
596 TREE_PREEMPT_RCU implementations, permitting Makefile to
597 trivially select kernel/rcutree_trace.c.
600 bool "Enable RCU priority boosting"
601 depends on RT_MUTEXES && PREEMPT_RCU
604 This option boosts the priority of preempted RCU readers that
605 block the current preemptible RCU grace period for too long.
606 This option also prevents heavy loads from blocking RCU
607 callback invocation for all flavors of RCU.
609 Say Y here if you are working with real-time apps or heavy loads
610 Say N here if you are unsure.
612 config RCU_BOOST_PRIO
613 int "Real-time priority to boost RCU readers to"
618 This option specifies the real-time priority to which long-term
619 preempted RCU readers are to be boosted. If you are working
620 with a real-time application that has one or more CPU-bound
621 threads running at a real-time priority level, you should set
622 RCU_BOOST_PRIO to a priority higher then the highest-priority
623 real-time CPU-bound thread. The default RCU_BOOST_PRIO value
624 of 1 is appropriate in the common case, which is real-time
625 applications that do not have any CPU-bound threads.
627 Some real-time applications might not have a single real-time
628 thread that saturates a given CPU, but instead might have
629 multiple real-time threads that, taken together, fully utilize
630 that CPU. In this case, you should set RCU_BOOST_PRIO to
631 a priority higher than the lowest-priority thread that is
632 conspiring to prevent the CPU from running any non-real-time
633 tasks. For example, if one thread at priority 10 and another
634 thread at priority 5 are between themselves fully consuming
635 the CPU time on a given CPU, then RCU_BOOST_PRIO should be
636 set to priority 6 or higher.
638 Specify the real-time priority, or take the default if unsure.
640 config RCU_BOOST_DELAY
641 int "Milliseconds to delay boosting after RCU grace-period start"
646 This option specifies the time to wait after the beginning of
647 a given grace period before priority-boosting preempted RCU
648 readers blocking that grace period. Note that any RCU reader
649 blocking an expedited RCU grace period is boosted immediately.
651 Accept the default if unsure.
654 bool "Offload RCU callback processing from boot-selected CPUs"
655 depends on TREE_RCU || TREE_PREEMPT_RCU
658 Use this option to reduce OS jitter for aggressive HPC or
659 real-time workloads. It can also be used to offload RCU
660 callback invocation to energy-efficient CPUs in battery-powered
661 asymmetric multiprocessors.
663 This option offloads callback invocation from the set of
664 CPUs specified at boot time by the rcu_nocbs parameter.
665 For each such CPU, a kthread ("rcuoN") will be created to
666 invoke callbacks, where the "N" is the CPU being offloaded.
667 Nothing prevents this kthread from running on the specified
668 CPUs, but (1) the kthreads may be preempted between each
669 callback, and (2) affinity or cgroups can be used to force
670 the kthreads to run on whatever set of CPUs is desired.
672 Say Y here if you want reduced OS jitter on selected CPUs.
673 Say N here if you are unsure.
675 endmenu # "RCU Subsystem"
678 tristate "Kernel .config support"
680 This option enables the complete Linux kernel ".config" file
681 contents to be saved in the kernel. It provides documentation
682 of which kernel options are used in a running kernel or in an
683 on-disk kernel. This information can be extracted from the kernel
684 image file with the script scripts/extract-ikconfig and used as
685 input to rebuild the current kernel or to build another kernel.
686 It can also be extracted from a running kernel by reading
687 /proc/config.gz if enabled (below).
690 bool "Enable access to .config through /proc/config.gz"
691 depends on IKCONFIG && PROC_FS
693 This option enables access to the kernel configuration file
694 through /proc/config.gz.
697 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
701 Select kernel log buffer size as a power of 2.
711 # Architectures with an unreliable sched_clock() should select this:
713 config HAVE_UNSTABLE_SCHED_CLOCK
717 # For architectures that want to enable the support for NUMA-affine scheduler
720 config ARCH_SUPPORTS_NUMA_BALANCING
723 # For architectures that (ab)use NUMA to represent different memory regions
724 # all cpu-local but of different latencies, such as SuperH.
726 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
730 # For architectures that are willing to define _PAGE_NUMA as _PAGE_PROTNONE
731 config ARCH_WANTS_PROT_NUMA_PROT_NONE
734 config ARCH_USES_NUMA_PROT_NONE
737 depends on ARCH_WANTS_PROT_NUMA_PROT_NONE
738 depends on NUMA_BALANCING
740 config NUMA_BALANCING_DEFAULT_ENABLED
741 bool "Automatically enable NUMA aware memory/task placement"
743 depends on NUMA_BALANCING
745 If set, autonumic NUMA balancing will be enabled if running on a NUMA
748 config NUMA_BALANCING
749 bool "Memory placement aware NUMA scheduler"
750 depends on ARCH_SUPPORTS_NUMA_BALANCING
751 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
752 depends on SMP && NUMA && MIGRATION
754 This option adds support for automatic NUMA aware memory/task placement.
755 The mechanism is quite primitive and is based on migrating memory when
756 it is references to the node the task is running on.
758 This system will be inactive on UMA systems.
761 boolean "Control Group support"
764 This option adds support for grouping sets of processes together, for
765 use with process control subsystems such as Cpusets, CFS, memory
766 controls or device isolation.
768 - Documentation/scheduler/sched-design-CFS.txt (CFS)
769 - Documentation/cgroups/ (features for grouping, isolation
770 and resource control)
777 bool "Example debug cgroup subsystem"
780 This option enables a simple cgroup subsystem that
781 exports useful debugging information about the cgroups
786 config CGROUP_FREEZER
787 bool "Freezer cgroup subsystem"
789 Provides a way to freeze and unfreeze all tasks in a
793 bool "Device controller for cgroups"
795 Provides a cgroup implementing whitelists for devices which
796 a process in the cgroup can mknod or open.
799 bool "Cpuset support"
801 This option will let you create and manage CPUSETs which
802 allow dynamically partitioning a system into sets of CPUs and
803 Memory Nodes and assigning tasks to run only within those sets.
804 This is primarily useful on large SMP or NUMA systems.
808 config PROC_PID_CPUSET
809 bool "Include legacy /proc/<pid>/cpuset file"
813 config CGROUP_CPUACCT
814 bool "Simple CPU accounting cgroup subsystem"
816 Provides a simple Resource Controller for monitoring the
817 total CPU consumed by the tasks in a cgroup.
819 config RESOURCE_COUNTERS
820 bool "Resource counters"
822 This option enables controller independent resource accounting
823 infrastructure that works with cgroups.
826 bool "Memory Resource Controller for Control Groups"
827 depends on RESOURCE_COUNTERS
830 Provides a memory resource controller that manages both anonymous
831 memory and page cache. (See Documentation/cgroups/memory.txt)
833 Note that setting this option increases fixed memory overhead
834 associated with each page of memory in the system. By this,
835 20(40)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory
836 usage tracking struct at boot. Total amount of this is printed out
839 Only enable when you're ok with these trade offs and really
840 sure you need the memory resource controller. Even when you enable
841 this, you can set "cgroup_disable=memory" at your boot option to
842 disable memory resource controller and you can avoid overheads.
843 (and lose benefits of memory resource controller)
845 This config option also selects MM_OWNER config option, which
846 could in turn add some fork/exit overhead.
849 bool "Memory Resource Controller Swap Extension"
850 depends on MEMCG && SWAP
852 Add swap management feature to memory resource controller. When you
853 enable this, you can limit mem+swap usage per cgroup. In other words,
854 when you disable this, memory resource controller has no cares to
855 usage of swap...a process can exhaust all of the swap. This extension
856 is useful when you want to avoid exhaustion swap but this itself
857 adds more overheads and consumes memory for remembering information.
858 Especially if you use 32bit system or small memory system, please
859 be careful about enabling this. When memory resource controller
860 is disabled by boot option, this will be automatically disabled and
861 there will be no overhead from this. Even when you set this config=y,
862 if boot option "swapaccount=0" is set, swap will not be accounted.
863 Now, memory usage of swap_cgroup is 2 bytes per entry. If swap page
864 size is 4096bytes, 512k per 1Gbytes of swap.
865 config MEMCG_SWAP_ENABLED
866 bool "Memory Resource Controller Swap Extension enabled by default"
867 depends on MEMCG_SWAP
870 Memory Resource Controller Swap Extension comes with its price in
871 a bigger memory consumption. General purpose distribution kernels
872 which want to enable the feature but keep it disabled by default
873 and let the user enable it by swapaccount boot command line
874 parameter should have this option unselected.
875 For those who want to have the feature enabled by default should
876 select this option (if, for some reason, they need to disable it
877 then swapaccount=0 does the trick).
879 bool "Memory Resource Controller Kernel Memory accounting (EXPERIMENTAL)"
880 depends on MEMCG && EXPERIMENTAL
881 depends on SLUB || SLAB
883 The Kernel Memory extension for Memory Resource Controller can limit
884 the amount of memory used by kernel objects in the system. Those are
885 fundamentally different from the entities handled by the standard
886 Memory Controller, which are page-based, and can be swapped. Users of
887 the kmem extension can use it to guarantee that no group of processes
888 will ever exhaust kernel resources alone.
890 config CGROUP_HUGETLB
891 bool "HugeTLB Resource Controller for Control Groups"
892 depends on RESOURCE_COUNTERS && HUGETLB_PAGE && EXPERIMENTAL
895 Provides a cgroup Resource Controller for HugeTLB pages.
896 When you enable this, you can put a per cgroup limit on HugeTLB usage.
897 The limit is enforced during page fault. Since HugeTLB doesn't
898 support page reclaim, enforcing the limit at page fault time implies
899 that, the application will get SIGBUS signal if it tries to access
900 HugeTLB pages beyond its limit. This requires the application to know
901 beforehand how much HugeTLB pages it would require for its use. The
902 control group is tracked in the third page lru pointer. This means
903 that we cannot use the controller with huge page less than 3 pages.
906 bool "Enable perf_event per-cpu per-container group (cgroup) monitoring"
907 depends on PERF_EVENTS && CGROUPS
909 This option extends the per-cpu mode to restrict monitoring to
910 threads which belong to the cgroup specified and run on the
915 menuconfig CGROUP_SCHED
916 bool "Group CPU scheduler"
919 This feature lets CPU scheduler recognize task groups and control CPU
920 bandwidth allocation to such task groups. It uses cgroups to group
924 config FAIR_GROUP_SCHED
925 bool "Group scheduling for SCHED_OTHER"
926 depends on CGROUP_SCHED
930 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
931 depends on EXPERIMENTAL
932 depends on FAIR_GROUP_SCHED
935 This option allows users to define CPU bandwidth rates (limits) for
936 tasks running within the fair group scheduler. Groups with no limit
937 set are considered to be unconstrained and will run with no
939 See tip/Documentation/scheduler/sched-bwc.txt for more information.
941 config RT_GROUP_SCHED
942 bool "Group scheduling for SCHED_RR/FIFO"
943 depends on EXPERIMENTAL
944 depends on CGROUP_SCHED
947 This feature lets you explicitly allocate real CPU bandwidth
948 to task groups. If enabled, it will also make it impossible to
949 schedule realtime tasks for non-root users until you allocate
950 realtime bandwidth for them.
951 See Documentation/scheduler/sched-rt-group.txt for more information.
956 bool "Block IO controller"
960 Generic block IO controller cgroup interface. This is the common
961 cgroup interface which should be used by various IO controlling
964 Currently, CFQ IO scheduler uses it to recognize task groups and
965 control disk bandwidth allocation (proportional time slice allocation)
966 to such task groups. It is also used by bio throttling logic in
967 block layer to implement upper limit in IO rates on a device.
969 This option only enables generic Block IO controller infrastructure.
970 One needs to also enable actual IO controlling logic/policy. For
971 enabling proportional weight division of disk bandwidth in CFQ, set
972 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
973 CONFIG_BLK_DEV_THROTTLING=y.
975 See Documentation/cgroups/blkio-controller.txt for more information.
977 config DEBUG_BLK_CGROUP
978 bool "Enable Block IO controller debugging"
979 depends on BLK_CGROUP
982 Enable some debugging help. Currently it exports additional stat
983 files in a cgroup which can be useful for debugging.
987 config CHECKPOINT_RESTORE
988 bool "Checkpoint/restore support" if EXPERT
991 Enables additional kernel features in a sake of checkpoint/restore.
992 In particular it adds auxiliary prctl codes to setup process text,
993 data and heap segment sizes, and a few additional /proc filesystem
996 If unsure, say N here.
998 menuconfig NAMESPACES
999 bool "Namespaces support" if EXPERT
1002 Provides the way to make tasks work with different objects using
1003 the same id. For example same IPC id may refer to different objects
1004 or same user id or pid may refer to different tasks when used in
1005 different namespaces.
1010 bool "UTS namespace"
1013 In this namespace tasks see different info provided with the
1017 bool "IPC namespace"
1018 depends on (SYSVIPC || POSIX_MQUEUE)
1021 In this namespace tasks work with IPC ids which correspond to
1022 different IPC objects in different namespaces.
1025 bool "User namespace (EXPERIMENTAL)"
1026 depends on EXPERIMENTAL
1027 depends on UIDGID_CONVERTED
1028 select UIDGID_STRICT_TYPE_CHECKS
1032 This allows containers, i.e. vservers, to use user namespaces
1033 to provide different user info for different servers.
1037 bool "PID Namespaces"
1040 Support process id namespaces. This allows having multiple
1041 processes with the same pid as long as they are in different
1042 pid namespaces. This is a building block of containers.
1045 bool "Network namespace"
1049 Allow user space to create what appear to be multiple instances
1050 of the network stack.
1054 config UIDGID_CONVERTED
1055 # True if all of the selected software conmponents are known
1056 # to have uid_t and gid_t converted to kuid_t and kgid_t
1057 # where appropriate and are otherwise safe to use with
1058 # the user namespace.
1063 depends on NET_9P = n
1066 depends on 9P_FS = n
1067 depends on AFS_FS = n
1068 depends on CEPH_FS = n
1070 depends on CODA_FS = n
1071 depends on GFS2_FS = n
1072 depends on NCP_FS = n
1074 depends on NFS_FS = n
1075 depends on OCFS2_FS = n
1076 depends on XFS_FS = n
1078 config UIDGID_STRICT_TYPE_CHECKS
1079 bool "Require conversions between uid/gids and their internal representation"
1080 depends on UIDGID_CONVERTED
1083 While the nececessary conversions are being added to all subsystems this option allows
1084 the code to continue to build for unconverted subsystems.
1086 Say Y here if you want the strict type checking enabled
1088 config SCHED_AUTOGROUP
1089 bool "Automatic process group scheduling"
1093 select FAIR_GROUP_SCHED
1095 This option optimizes the scheduler for common desktop workloads by
1096 automatically creating and populating task groups. This separation
1097 of workloads isolates aggressive CPU burners (like build jobs) from
1098 desktop applications. Task group autogeneration is currently based
1104 config SYSFS_DEPRECATED
1105 bool "Enable deprecated sysfs features to support old userspace tools"
1109 This option adds code that switches the layout of the "block" class
1110 devices, to not show up in /sys/class/block/, but only in
1113 This switch is only active when the sysfs.deprecated=1 boot option is
1114 passed or the SYSFS_DEPRECATED_V2 option is set.
1116 This option allows new kernels to run on old distributions and tools,
1117 which might get confused by /sys/class/block/. Since 2007/2008 all
1118 major distributions and tools handle this just fine.
1120 Recent distributions and userspace tools after 2009/2010 depend on
1121 the existence of /sys/class/block/, and will not work with this
1124 Only if you are using a new kernel on an old distribution, you might
1127 config SYSFS_DEPRECATED_V2
1128 bool "Enable deprecated sysfs features by default"
1131 depends on SYSFS_DEPRECATED
1133 Enable deprecated sysfs by default.
1135 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1138 Only if you are using a new kernel on an old distribution, you might
1139 need to say Y here. Even then, odds are you would not need it
1140 enabled, you can always pass the boot option if absolutely necessary.
1143 bool "Kernel->user space relay support (formerly relayfs)"
1145 This option enables support for relay interface support in
1146 certain file systems (such as debugfs).
1147 It is designed to provide an efficient mechanism for tools and
1148 facilities to relay large amounts of data from kernel space to
1153 config BLK_DEV_INITRD
1154 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1155 depends on BROKEN || !FRV
1157 The initial RAM filesystem is a ramfs which is loaded by the
1158 boot loader (loadlin or lilo) and that is mounted as root
1159 before the normal boot procedure. It is typically used to
1160 load modules needed to mount the "real" root file system,
1161 etc. See <file:Documentation/initrd.txt> for details.
1163 If RAM disk support (BLK_DEV_RAM) is also included, this
1164 also enables initial RAM disk (initrd) support and adds
1165 15 Kbytes (more on some other architectures) to the kernel size.
1171 source "usr/Kconfig"
1175 config CC_OPTIMIZE_FOR_SIZE
1176 bool "Optimize for size"
1178 Enabling this option will pass "-Os" instead of "-O2" to gcc
1179 resulting in a smaller kernel.
1190 bool "Configure standard kernel features (expert users)"
1191 # Unhide debug options, to make the on-by-default options visible
1194 This option allows certain base kernel options and settings
1195 to be disabled or tweaked. This is for specialized
1196 environments which can tolerate a "non-standard" kernel.
1197 Only use this if you really know what you are doing.
1203 bool "Enable 16-bit UID system calls" if EXPERT
1204 depends on HAVE_UID16
1207 This enables the legacy 16-bit UID syscall wrappers.
1209 config SYSCTL_SYSCALL
1210 bool "Sysctl syscall support" if EXPERT
1211 depends on PROC_SYSCTL
1215 sys_sysctl uses binary paths that have been found challenging
1216 to properly maintain and use. The interface in /proc/sys
1217 using paths with ascii names is now the primary path to this
1220 Almost nothing using the binary sysctl interface so if you are
1221 trying to save some space it is probably safe to disable this,
1222 making your kernel marginally smaller.
1224 If unsure say N here.
1226 config SYSCTL_EXCEPTION_TRACE
1229 Enable support for /proc/sys/debug/exception-trace.
1232 bool "Load all symbols for debugging/ksymoops" if EXPERT
1235 Say Y here to let the kernel print out symbolic crash information and
1236 symbolic stack backtraces. This increases the size of the kernel
1237 somewhat, as all symbols have to be loaded into the kernel image.
1240 bool "Include all symbols in kallsyms"
1241 depends on DEBUG_KERNEL && KALLSYMS
1243 Normally kallsyms only contains the symbols of functions for nicer
1244 OOPS messages and backtraces (i.e., symbols from the text and inittext
1245 sections). This is sufficient for most cases. And only in very rare
1246 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1247 names of variables from the data sections, etc).
1249 This option makes sure that all symbols are loaded into the kernel
1250 image (i.e., symbols from all sections) in cost of increased kernel
1251 size (depending on the kernel configuration, it may be 300KiB or
1252 something like this).
1254 Say N unless you really need all symbols.
1261 bool "Enable support for printk" if EXPERT
1264 This option enables normal printk support. Removing it
1265 eliminates most of the message strings from the kernel image
1266 and makes the kernel more or less silent. As this makes it
1267 very difficult to diagnose system problems, saying N here is
1268 strongly discouraged.
1271 bool "BUG() support" if EXPERT
1274 Disabling this option eliminates support for BUG and WARN, reducing
1275 the size of your kernel image and potentially quietly ignoring
1276 numerous fatal conditions. You should only consider disabling this
1277 option for embedded systems with no facilities for reporting errors.
1283 bool "Enable ELF core dumps" if EXPERT
1285 Enable support for generating core dumps. Disabling saves about 4k.
1288 config PCSPKR_PLATFORM
1289 bool "Enable PC-Speaker support" if EXPERT
1290 depends on HAVE_PCSPKR_PLATFORM
1294 This option allows to disable the internal PC-Speaker
1295 support, saving some memory.
1297 config HAVE_PCSPKR_PLATFORM
1302 bool "Enable full-sized data structures for core" if EXPERT
1304 Disabling this option reduces the size of miscellaneous core
1305 kernel data structures. This saves memory on small machines,
1306 but may reduce performance.
1309 bool "Enable futex support" if EXPERT
1313 Disabling this option will cause the kernel to be built without
1314 support for "fast userspace mutexes". The resulting kernel may not
1315 run glibc-based applications correctly.
1318 bool "Enable eventpoll support" if EXPERT
1322 Disabling this option will cause the kernel to be built without
1323 support for epoll family of system calls.
1326 bool "Enable signalfd() system call" if EXPERT
1330 Enable the signalfd() system call that allows to receive signals
1331 on a file descriptor.
1336 bool "Enable timerfd() system call" if EXPERT
1340 Enable the timerfd() system call that allows to receive timer
1341 events on a file descriptor.
1346 bool "Enable eventfd() system call" if EXPERT
1350 Enable the eventfd() system call that allows to receive both
1351 kernel notification (ie. KAIO) or userspace notifications.
1356 bool "Use full shmem filesystem" if EXPERT
1360 The shmem is an internal filesystem used to manage shared memory.
1361 It is backed by swap and manages resource limits. It is also exported
1362 to userspace as tmpfs if TMPFS is enabled. Disabling this
1363 option replaces shmem and tmpfs with the much simpler ramfs code,
1364 which may be appropriate on small systems without swap.
1367 bool "Enable AIO support" if EXPERT
1370 This option enables POSIX asynchronous I/O which may by used
1371 by some high performance threaded applications. Disabling
1372 this option saves about 7k.
1375 bool "Embedded system"
1378 This option should be enabled if compiling the kernel for
1379 an embedded system so certain expert options are available
1382 config HAVE_PERF_EVENTS
1385 See tools/perf/design.txt for details.
1387 config PERF_USE_VMALLOC
1390 See tools/perf/design.txt for details
1392 menu "Kernel Performance Events And Counters"
1395 bool "Kernel performance events and counters"
1396 default y if PROFILING
1397 depends on HAVE_PERF_EVENTS
1401 Enable kernel support for various performance events provided
1402 by software and hardware.
1404 Software events are supported either built-in or via the
1405 use of generic tracepoints.
1407 Most modern CPUs support performance events via performance
1408 counter registers. These registers count the number of certain
1409 types of hw events: such as instructions executed, cachemisses
1410 suffered, or branches mis-predicted - without slowing down the
1411 kernel or applications. These registers can also trigger interrupts
1412 when a threshold number of events have passed - and can thus be
1413 used to profile the code that runs on that CPU.
1415 The Linux Performance Event subsystem provides an abstraction of
1416 these software and hardware event capabilities, available via a
1417 system call and used by the "perf" utility in tools/perf/. It
1418 provides per task and per CPU counters, and it provides event
1419 capabilities on top of those.
1423 config DEBUG_PERF_USE_VMALLOC
1425 bool "Debug: use vmalloc to back perf mmap() buffers"
1426 depends on PERF_EVENTS && DEBUG_KERNEL
1427 select PERF_USE_VMALLOC
1429 Use vmalloc memory to back perf mmap() buffers.
1431 Mostly useful for debugging the vmalloc code on platforms
1432 that don't require it.
1438 config VM_EVENT_COUNTERS
1440 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1442 VM event counters are needed for event counts to be shown.
1443 This option allows the disabling of the VM event counters
1444 on EXPERT systems. /proc/vmstat will only show page counts
1445 if VM event counters are disabled.
1449 bool "Enable PCI quirk workarounds" if EXPERT
1452 This enables workarounds for various PCI chipset
1453 bugs/quirks. Disable this only if your target machine is
1454 unaffected by PCI quirks.
1458 bool "Enable SLUB debugging support" if EXPERT
1459 depends on SLUB && SYSFS
1461 SLUB has extensive debug support features. Disabling these can
1462 result in significant savings in code size. This also disables
1463 SLUB sysfs support. /sys/slab will not exist and there will be
1464 no support for cache validation etc.
1467 bool "Disable heap randomization"
1470 Randomizing heap placement makes heap exploits harder, but it
1471 also breaks ancient binaries (including anything libc5 based).
1472 This option changes the bootup default to heap randomization
1473 disabled, and can be overridden at runtime by setting
1474 /proc/sys/kernel/randomize_va_space to 2.
1476 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1479 prompt "Choose SLAB allocator"
1482 This option allows to select a slab allocator.
1487 The regular slab allocator that is established and known to work
1488 well in all environments. It organizes cache hot objects in
1489 per cpu and per node queues.
1492 bool "SLUB (Unqueued Allocator)"
1494 SLUB is a slab allocator that minimizes cache line usage
1495 instead of managing queues of cached objects (SLAB approach).
1496 Per cpu caching is realized using slabs of objects instead
1497 of queues of objects. SLUB can use memory efficiently
1498 and has enhanced diagnostics. SLUB is the default choice for
1503 bool "SLOB (Simple Allocator)"
1505 SLOB replaces the stock allocator with a drastically simpler
1506 allocator. SLOB is generally more space efficient but
1507 does not perform as well on large systems.
1511 config MMAP_ALLOW_UNINITIALIZED
1512 bool "Allow mmapped anonymous memory to be uninitialized"
1513 depends on EXPERT && !MMU
1516 Normally, and according to the Linux spec, anonymous memory obtained
1517 from mmap() has it's contents cleared before it is passed to
1518 userspace. Enabling this config option allows you to request that
1519 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1520 providing a huge performance boost. If this option is not enabled,
1521 then the flag will be ignored.
1523 This is taken advantage of by uClibc's malloc(), and also by
1524 ELF-FDPIC binfmt's brk and stack allocator.
1526 Because of the obvious security issues, this option should only be
1527 enabled on embedded devices where you control what is run in
1528 userspace. Since that isn't generally a problem on no-MMU systems,
1529 it is normally safe to say Y here.
1531 See Documentation/nommu-mmap.txt for more information.
1534 bool "Profiling support"
1536 Say Y here to enable the extended profiling support mechanisms used
1537 by profilers such as OProfile.
1540 # Place an empty function call at each tracepoint site. Can be
1541 # dynamically changed for a probe function.
1546 source "arch/Kconfig"
1548 endmenu # General setup
1550 config HAVE_GENERIC_DMA_COHERENT
1557 depends on SLAB || SLUB_DEBUG
1565 default 0 if BASE_FULL
1566 default 1 if !BASE_FULL
1569 bool "Enable loadable module support"
1571 Kernel modules are small pieces of compiled code which can
1572 be inserted in the running kernel, rather than being
1573 permanently built into the kernel. You use the "modprobe"
1574 tool to add (and sometimes remove) them. If you say Y here,
1575 many parts of the kernel can be built as modules (by
1576 answering M instead of Y where indicated): this is most
1577 useful for infrequently used options which are not required
1578 for booting. For more information, see the man pages for
1579 modprobe, lsmod, modinfo, insmod and rmmod.
1581 If you say Y here, you will need to run "make
1582 modules_install" to put the modules under /lib/modules/
1583 where modprobe can find them (you may need to be root to do
1590 config MODULE_FORCE_LOAD
1591 bool "Forced module loading"
1594 Allow loading of modules without version information (ie. modprobe
1595 --force). Forced module loading sets the 'F' (forced) taint flag and
1596 is usually a really bad idea.
1598 config MODULE_UNLOAD
1599 bool "Module unloading"
1601 Without this option you will not be able to unload any
1602 modules (note that some modules may not be unloadable
1603 anyway), which makes your kernel smaller, faster
1604 and simpler. If unsure, say Y.
1606 config MODULE_FORCE_UNLOAD
1607 bool "Forced module unloading"
1608 depends on MODULE_UNLOAD && EXPERIMENTAL
1610 This option allows you to force a module to unload, even if the
1611 kernel believes it is unsafe: the kernel will remove the module
1612 without waiting for anyone to stop using it (using the -f option to
1613 rmmod). This is mainly for kernel developers and desperate users.
1617 bool "Module versioning support"
1619 Usually, you have to use modules compiled with your kernel.
1620 Saying Y here makes it sometimes possible to use modules
1621 compiled for different kernels, by adding enough information
1622 to the modules to (hopefully) spot any changes which would
1623 make them incompatible with the kernel you are running. If
1626 config MODULE_SRCVERSION_ALL
1627 bool "Source checksum for all modules"
1629 Modules which contain a MODULE_VERSION get an extra "srcversion"
1630 field inserted into their modinfo section, which contains a
1631 sum of the source files which made it. This helps maintainers
1632 see exactly which source was used to build a module (since
1633 others sometimes change the module source without updating
1634 the version). With this option, such a "srcversion" field
1635 will be created for all modules. If unsure, say N.
1638 bool "Module signature verification"
1642 select ASYMMETRIC_KEY_TYPE
1643 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1644 select PUBLIC_KEY_ALGO_RSA
1647 select X509_CERTIFICATE_PARSER
1649 Check modules for valid signatures upon load: the signature
1650 is simply appended to the module. For more information see
1651 Documentation/module-signing.txt.
1653 !!!WARNING!!! If you enable this option, you MUST make sure that the
1654 module DOES NOT get stripped after being signed. This includes the
1655 debuginfo strip done by some packagers (such as rpmbuild) and
1656 inclusion into an initramfs that wants the module size reduced.
1658 config MODULE_SIG_FORCE
1659 bool "Require modules to be validly signed"
1660 depends on MODULE_SIG
1662 Reject unsigned modules or signed modules for which we don't have a
1663 key. Without this, such modules will simply taint the kernel.
1666 prompt "Which hash algorithm should modules be signed with?"
1667 depends on MODULE_SIG
1669 This determines which sort of hashing algorithm will be used during
1670 signature generation. This algorithm _must_ be built into the kernel
1671 directly so that signature verification can take place. It is not
1672 possible to load a signed module containing the algorithm to check
1673 the signature on that module.
1675 config MODULE_SIG_SHA1
1676 bool "Sign modules with SHA-1"
1679 config MODULE_SIG_SHA224
1680 bool "Sign modules with SHA-224"
1681 select CRYPTO_SHA256
1683 config MODULE_SIG_SHA256
1684 bool "Sign modules with SHA-256"
1685 select CRYPTO_SHA256
1687 config MODULE_SIG_SHA384
1688 bool "Sign modules with SHA-384"
1689 select CRYPTO_SHA512
1691 config MODULE_SIG_SHA512
1692 bool "Sign modules with SHA-512"
1693 select CRYPTO_SHA512
1699 config INIT_ALL_POSSIBLE
1702 Back when each arch used to define their own cpu_online_mask and
1703 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
1704 with all 1s, and others with all 0s. When they were centralised,
1705 it was better to provide this option than to break all the archs
1706 and have several arch maintainers pursuing me down dark alleys.
1711 depends on (SMP && MODULE_UNLOAD) || HOTPLUG_CPU
1713 Need stop_machine() primitive.
1715 source "block/Kconfig"
1717 config PREEMPT_NOTIFIERS
1724 # Can be selected by architectures with broken toolchains
1725 # that get confused by correct const<->read_only section
1727 config BROKEN_RODATA
1733 Build a simple ASN.1 grammar compiler that produces a bytecode output
1734 that can be interpreted by the ASN.1 stream decoder and used to
1735 inform it as to what tags are to be expected in a stream and what
1736 functions to call on what tags.
1738 source "kernel/Kconfig.locks"