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"
28 depends on HAVE_IRQ_WORK
33 bool "Prompt for development and/or incomplete code/drivers"
35 Some of the various things that Linux supports (such as network
36 drivers, file systems, network protocols, etc.) can be in a state
37 of development where the functionality, stability, or the level of
38 testing is not yet high enough for general use. This is usually
39 known as the "alpha-test" phase among developers. If a feature is
40 currently in alpha-test, then the developers usually discourage
41 uninformed widespread use of this feature by the general public to
42 avoid "Why doesn't this work?" type mail messages. However, active
43 testing and use of these systems is welcomed. Just be aware that it
44 may not meet the normal level of reliability or it may fail to work
45 in some special cases. Detailed bug reports from people familiar
46 with the kernel internals are usually welcomed by the developers
47 (before submitting bug reports, please read the documents
48 <file:README>, <file:MAINTAINERS>, <file:REPORTING-BUGS>,
49 <file:Documentation/BUG-HUNTING>, and
50 <file:Documentation/oops-tracing.txt> in the kernel source).
52 This option will also make obsoleted drivers available. These are
53 drivers that have been replaced by something else, and/or are
54 scheduled to be removed in a future kernel release.
56 Unless you intend to help test and develop a feature or driver that
57 falls into this category, or you have a situation that requires
58 using these features, you should probably say N here, which will
59 cause the configurator to present you with fewer choices. If
60 you say Y here, you will be offered the choice of using features or
61 drivers that are currently considered to be in the alpha-test phase.
68 depends on BROKEN || !SMP
71 config INIT_ENV_ARG_LIMIT
76 Maximum of each of the number of arguments and environment
77 variables passed to init from the kernel command line.
81 string "Cross-compiler tool prefix"
83 Same as running 'make CROSS_COMPILE=prefix-' but stored for
84 default make runs in this kernel build directory. You don't
85 need to set this unless you want the configured kernel build
86 directory to select the cross-compiler automatically.
89 string "Local version - append to kernel release"
91 Append an extra string to the end of your kernel version.
92 This will show up when you type uname, for example.
93 The string you set here will be appended after the contents of
94 any files with a filename matching localversion* in your
95 object and source tree, in that order. Your total string can
96 be a maximum of 64 characters.
98 config LOCALVERSION_AUTO
99 bool "Automatically append version information to the version string"
102 This will try to automatically determine if the current tree is a
103 release tree by looking for git tags that belong to the current
104 top of tree revision.
106 A string of the format -gxxxxxxxx will be added to the localversion
107 if a git-based tree is found. The string generated by this will be
108 appended after any matching localversion* files, and after the value
109 set in CONFIG_LOCALVERSION.
111 (The actual string used here is the first eight characters produced
112 by running the command:
114 $ git rev-parse --verify HEAD
116 which is done within the script "scripts/setlocalversion".)
118 config HAVE_KERNEL_GZIP
121 config HAVE_KERNEL_BZIP2
124 config HAVE_KERNEL_LZMA
127 config HAVE_KERNEL_XZ
130 config HAVE_KERNEL_LZO
134 prompt "Kernel compression mode"
136 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO
138 The linux kernel is a kind of self-extracting executable.
139 Several compression algorithms are available, which differ
140 in efficiency, compression and decompression speed.
141 Compression speed is only relevant when building a kernel.
142 Decompression speed is relevant at each boot.
144 If you have any problems with bzip2 or lzma compressed
145 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
146 version of this functionality (bzip2 only), for 2.4, was
147 supplied by Christian Ludwig)
149 High compression options are mostly useful for users, who
150 are low on disk space (embedded systems), but for whom ram
153 If in doubt, select 'gzip'
157 depends on HAVE_KERNEL_GZIP
159 The old and tried gzip compression. It provides a good balance
160 between compression ratio and decompression speed.
164 depends on HAVE_KERNEL_BZIP2
166 Its compression ratio and speed is intermediate.
167 Decompression speed is slowest among the three. The kernel
168 size is about 10% smaller with bzip2, in comparison to gzip.
169 Bzip2 uses a large amount of memory. For modern kernels you
170 will need at least 8MB RAM or more for booting.
174 depends on HAVE_KERNEL_LZMA
176 The most recent compression algorithm.
177 Its ratio is best, decompression speed is between the other
178 two. Compression is slowest. The kernel size is about 33%
179 smaller with LZMA in comparison to gzip.
183 depends on HAVE_KERNEL_XZ
185 XZ uses the LZMA2 algorithm and instruction set specific
186 BCJ filters which can improve compression ratio of executable
187 code. The size of the kernel is about 30% smaller with XZ in
188 comparison to gzip. On architectures for which there is a BCJ
189 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
190 will create a few percent smaller kernel than plain LZMA.
192 The speed is about the same as with LZMA: The decompression
193 speed of XZ is better than that of bzip2 but worse than gzip
194 and LZO. Compression is slow.
198 depends on HAVE_KERNEL_LZO
200 Its compression ratio is the poorest among the 4. The kernel
201 size is about 10% bigger than gzip; however its speed
202 (both compression and decompression) is the fastest.
206 config DEFAULT_HOSTNAME
207 string "Default hostname"
210 This option determines the default system hostname before userspace
211 calls sethostname(2). The kernel traditionally uses "(none)" here,
212 but you may wish to use a different default here to make a minimal
213 system more usable with less configuration.
216 bool "Support for paging of anonymous memory (swap)"
217 depends on MMU && BLOCK
220 This option allows you to choose whether you want to have support
221 for so called swap devices or swap files in your kernel that are
222 used to provide more virtual memory than the actual RAM present
223 in your computer. If unsure say Y.
228 Inter Process Communication is a suite of library functions and
229 system calls which let processes (running programs) synchronize and
230 exchange information. It is generally considered to be a good thing,
231 and some programs won't run unless you say Y here. In particular, if
232 you want to run the DOS emulator dosemu under Linux (read the
233 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
234 you'll need to say Y here.
236 You can find documentation about IPC with "info ipc" and also in
237 section 6.4 of the Linux Programmer's Guide, available from
238 <http://www.tldp.org/guides.html>.
240 config SYSVIPC_SYSCTL
247 bool "POSIX Message Queues"
248 depends on NET && EXPERIMENTAL
250 POSIX variant of message queues is a part of IPC. In POSIX message
251 queues every message has a priority which decides about succession
252 of receiving it by a process. If you want to compile and run
253 programs written e.g. for Solaris with use of its POSIX message
254 queues (functions mq_*) say Y here.
256 POSIX message queues are visible as a filesystem called 'mqueue'
257 and can be mounted somewhere if you want to do filesystem
258 operations on message queues.
262 config POSIX_MQUEUE_SYSCTL
264 depends on POSIX_MQUEUE
268 config BSD_PROCESS_ACCT
269 bool "BSD Process Accounting"
271 If you say Y here, a user level program will be able to instruct the
272 kernel (via a special system call) to write process accounting
273 information to a file: whenever a process exits, information about
274 that process will be appended to the file by the kernel. The
275 information includes things such as creation time, owning user,
276 command name, memory usage, controlling terminal etc. (the complete
277 list is in the struct acct in <file:include/linux/acct.h>). It is
278 up to the user level program to do useful things with this
279 information. This is generally a good idea, so say Y.
281 config BSD_PROCESS_ACCT_V3
282 bool "BSD Process Accounting version 3 file format"
283 depends on BSD_PROCESS_ACCT
286 If you say Y here, the process accounting information is written
287 in a new file format that also logs the process IDs of each
288 process and it's parent. Note that this file format is incompatible
289 with previous v0/v1/v2 file formats, so you will need updated tools
290 for processing it. A preliminary version of these tools is available
291 at <http://www.gnu.org/software/acct/>.
294 bool "open by fhandle syscalls"
297 If you say Y here, a user level program will be able to map
298 file names to handle and then later use the handle for
299 different file system operations. This is useful in implementing
300 userspace file servers, which now track files using handles instead
301 of names. The handle would remain the same even if file names
302 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
306 bool "Export task/process statistics through netlink (EXPERIMENTAL)"
310 Export selected statistics for tasks/processes through the
311 generic netlink interface. Unlike BSD process accounting, the
312 statistics are available during the lifetime of tasks/processes as
313 responses to commands. Like BSD accounting, they are sent to user
318 config TASK_DELAY_ACCT
319 bool "Enable per-task delay accounting (EXPERIMENTAL)"
322 Collect information on time spent by a task waiting for system
323 resources like cpu, synchronous block I/O completion and swapping
324 in pages. Such statistics can help in setting a task's priorities
325 relative to other tasks for cpu, io, rss limits etc.
330 bool "Enable extended accounting over taskstats (EXPERIMENTAL)"
333 Collect extended task accounting data and send the data
334 to userland for processing over the taskstats interface.
338 config TASK_IO_ACCOUNTING
339 bool "Enable per-task storage I/O accounting (EXPERIMENTAL)"
340 depends on TASK_XACCT
342 Collect information on the number of bytes of storage I/O which this
348 bool "Auditing support"
351 Enable auditing infrastructure that can be used with another
352 kernel subsystem, such as SELinux (which requires this for
353 logging of avc messages output). Does not do system-call
354 auditing without CONFIG_AUDITSYSCALL.
357 bool "Enable system-call auditing support"
358 depends on AUDIT && (X86 || PPC || S390 || IA64 || UML || SPARC64 || SUPERH || ARM)
359 default y if SECURITY_SELINUX
361 Enable low-overhead system-call auditing infrastructure that
362 can be used independently or with another kernel subsystem,
367 depends on AUDITSYSCALL
372 depends on AUDITSYSCALL
375 config AUDIT_LOGINUID_IMMUTABLE
376 bool "Make audit loginuid immutable"
379 The config option toggles if a task setting its loginuid requires
380 CAP_SYS_AUDITCONTROL or if that task should require no special permissions
381 but should instead only allow setting its loginuid if it was never
382 previously set. On systems which use systemd or a similar central
383 process to restart login services this should be set to true. On older
384 systems in which an admin would typically have to directly stop and
385 start processes this should be set to false. Setting this to true allows
386 one to drop potentially dangerous capabilites from the login tasks,
387 but may not be backwards compatible with older init systems.
389 source "kernel/irq/Kconfig"
394 prompt "RCU Implementation"
398 bool "Tree-based hierarchical RCU"
399 depends on !PREEMPT && SMP
401 This option selects the RCU implementation that is
402 designed for very large SMP system with hundreds or
403 thousands of CPUs. It also scales down nicely to
406 config TREE_PREEMPT_RCU
407 bool "Preemptible tree-based hierarchical RCU"
408 depends on PREEMPT && SMP
410 This option selects the RCU implementation that is
411 designed for very large SMP systems with hundreds or
412 thousands of CPUs, but for which real-time response
413 is also required. It also scales down nicely to
417 bool "UP-only small-memory-footprint RCU"
418 depends on !PREEMPT && !SMP
420 This option selects the RCU implementation that is
421 designed for UP systems from which real-time response
422 is not required. This option greatly reduces the
423 memory footprint of RCU.
425 config TINY_PREEMPT_RCU
426 bool "Preemptible UP-only small-memory-footprint RCU"
427 depends on PREEMPT && !SMP
429 This option selects the RCU implementation that is designed
430 for real-time UP systems. This option greatly reduces the
431 memory footprint of RCU.
436 def_bool ( TREE_PREEMPT_RCU || TINY_PREEMPT_RCU )
438 This option enables preemptible-RCU code that is common between
439 the TREE_PREEMPT_RCU and TINY_PREEMPT_RCU implementations.
442 int "Tree-based hierarchical RCU fanout value"
445 depends on TREE_RCU || TREE_PREEMPT_RCU
449 This option controls the fanout of hierarchical implementations
450 of RCU, allowing RCU to work efficiently on machines with
451 large numbers of CPUs. This value must be at least the fourth
452 root of NR_CPUS, which allows NR_CPUS to be insanely large.
453 The default value of RCU_FANOUT should be used for production
454 systems, but if you are stress-testing the RCU implementation
455 itself, small RCU_FANOUT values allow you to test large-system
456 code paths on small(er) systems.
458 Select a specific number if testing RCU itself.
459 Take the default if unsure.
461 config RCU_FANOUT_LEAF
462 int "Tree-based hierarchical RCU leaf-level fanout value"
463 range 2 RCU_FANOUT if 64BIT
464 range 2 RCU_FANOUT if !64BIT
465 depends on TREE_RCU || TREE_PREEMPT_RCU
468 This option controls the leaf-level fanout of hierarchical
469 implementations of RCU, and allows trading off cache misses
470 against lock contention. Systems that synchronize their
471 scheduling-clock interrupts for energy-efficiency reasons will
472 want the default because the smaller leaf-level fanout keeps
473 lock contention levels acceptably low. Very large systems
474 (hundreds or thousands of CPUs) will instead want to set this
475 value to the maximum value possible in order to reduce the
476 number of cache misses incurred during RCU's grace-period
477 initialization. These systems tend to run CPU-bound, and thus
478 are not helped by synchronized interrupts, and thus tend to
479 skew them, which reduces lock contention enough that large
480 leaf-level fanouts work well.
482 Select a specific number if testing RCU itself.
484 Select the maximum permissible value for large systems.
486 Take the default if unsure.
488 config RCU_FANOUT_EXACT
489 bool "Disable tree-based hierarchical RCU auto-balancing"
490 depends on TREE_RCU || TREE_PREEMPT_RCU
493 This option forces use of the exact RCU_FANOUT value specified,
494 regardless of imbalances in the hierarchy. This is useful for
495 testing RCU itself, and might one day be useful on systems with
496 strong NUMA behavior.
498 Without RCU_FANOUT_EXACT, the code will balance the hierarchy.
502 config RCU_FAST_NO_HZ
503 bool "Accelerate last non-dyntick-idle CPU's grace periods"
504 depends on NO_HZ && SMP
507 This option causes RCU to attempt to accelerate grace periods
508 in order to allow CPUs to enter dynticks-idle state more
509 quickly. On the other hand, this option increases the overhead
510 of the dynticks-idle checking, particularly on systems with
511 large numbers of CPUs.
513 Say Y if energy efficiency is critically important, particularly
514 if you have relatively few CPUs.
516 Say N if you are unsure.
518 config TREE_RCU_TRACE
519 def_bool RCU_TRACE && ( TREE_RCU || TREE_PREEMPT_RCU )
522 This option provides tracing for the TREE_RCU and
523 TREE_PREEMPT_RCU implementations, permitting Makefile to
524 trivially select kernel/rcutree_trace.c.
527 bool "Enable RCU priority boosting"
528 depends on RT_MUTEXES && PREEMPT_RCU
531 This option boosts the priority of preempted RCU readers that
532 block the current preemptible RCU grace period for too long.
533 This option also prevents heavy loads from blocking RCU
534 callback invocation for all flavors of RCU.
536 Say Y here if you are working with real-time apps or heavy loads
537 Say N here if you are unsure.
539 config RCU_BOOST_PRIO
540 int "Real-time priority to boost RCU readers to"
545 This option specifies the real-time priority to which long-term
546 preempted RCU readers are to be boosted. If you are working
547 with a real-time application that has one or more CPU-bound
548 threads running at a real-time priority level, you should set
549 RCU_BOOST_PRIO to a priority higher then the highest-priority
550 real-time CPU-bound thread. The default RCU_BOOST_PRIO value
551 of 1 is appropriate in the common case, which is real-time
552 applications that do not have any CPU-bound threads.
554 Some real-time applications might not have a single real-time
555 thread that saturates a given CPU, but instead might have
556 multiple real-time threads that, taken together, fully utilize
557 that CPU. In this case, you should set RCU_BOOST_PRIO to
558 a priority higher than the lowest-priority thread that is
559 conspiring to prevent the CPU from running any non-real-time
560 tasks. For example, if one thread at priority 10 and another
561 thread at priority 5 are between themselves fully consuming
562 the CPU time on a given CPU, then RCU_BOOST_PRIO should be
563 set to priority 6 or higher.
565 Specify the real-time priority, or take the default if unsure.
567 config RCU_BOOST_DELAY
568 int "Milliseconds to delay boosting after RCU grace-period start"
573 This option specifies the time to wait after the beginning of
574 a given grace period before priority-boosting preempted RCU
575 readers blocking that grace period. Note that any RCU reader
576 blocking an expedited RCU grace period is boosted immediately.
578 Accept the default if unsure.
580 endmenu # "RCU Subsystem"
583 tristate "Kernel .config support"
585 This option enables the complete Linux kernel ".config" file
586 contents to be saved in the kernel. It provides documentation
587 of which kernel options are used in a running kernel or in an
588 on-disk kernel. This information can be extracted from the kernel
589 image file with the script scripts/extract-ikconfig and used as
590 input to rebuild the current kernel or to build another kernel.
591 It can also be extracted from a running kernel by reading
592 /proc/config.gz if enabled (below).
595 bool "Enable access to .config through /proc/config.gz"
596 depends on IKCONFIG && PROC_FS
598 This option enables access to the kernel configuration file
599 through /proc/config.gz.
602 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
606 Select kernel log buffer size as a power of 2.
616 # Architectures with an unreliable sched_clock() should select this:
618 config HAVE_UNSTABLE_SCHED_CLOCK
622 boolean "Control Group support"
625 This option adds support for grouping sets of processes together, for
626 use with process control subsystems such as Cpusets, CFS, memory
627 controls or device isolation.
629 - Documentation/scheduler/sched-design-CFS.txt (CFS)
630 - Documentation/cgroups/ (features for grouping, isolation
631 and resource control)
638 bool "Example debug cgroup subsystem"
641 This option enables a simple cgroup subsystem that
642 exports useful debugging information about the cgroups
647 config CGROUP_FREEZER
648 bool "Freezer cgroup subsystem"
650 Provides a way to freeze and unfreeze all tasks in a
654 bool "Device controller for cgroups"
656 Provides a cgroup implementing whitelists for devices which
657 a process in the cgroup can mknod or open.
660 bool "Cpuset support"
662 This option will let you create and manage CPUSETs which
663 allow dynamically partitioning a system into sets of CPUs and
664 Memory Nodes and assigning tasks to run only within those sets.
665 This is primarily useful on large SMP or NUMA systems.
669 config PROC_PID_CPUSET
670 bool "Include legacy /proc/<pid>/cpuset file"
674 config CGROUP_CPUACCT
675 bool "Simple CPU accounting cgroup subsystem"
677 Provides a simple Resource Controller for monitoring the
678 total CPU consumed by the tasks in a cgroup.
680 config RESOURCE_COUNTERS
681 bool "Resource counters"
683 This option enables controller independent resource accounting
684 infrastructure that works with cgroups.
686 config CGROUP_MEM_RES_CTLR
687 bool "Memory Resource Controller for Control Groups"
688 depends on RESOURCE_COUNTERS
691 Provides a memory resource controller that manages both anonymous
692 memory and page cache. (See Documentation/cgroups/memory.txt)
694 Note that setting this option increases fixed memory overhead
695 associated with each page of memory in the system. By this,
696 20(40)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory
697 usage tracking struct at boot. Total amount of this is printed out
700 Only enable when you're ok with these trade offs and really
701 sure you need the memory resource controller. Even when you enable
702 this, you can set "cgroup_disable=memory" at your boot option to
703 disable memory resource controller and you can avoid overheads.
704 (and lose benefits of memory resource controller)
706 This config option also selects MM_OWNER config option, which
707 could in turn add some fork/exit overhead.
709 config CGROUP_MEM_RES_CTLR_SWAP
710 bool "Memory Resource Controller Swap Extension"
711 depends on CGROUP_MEM_RES_CTLR && SWAP
713 Add swap management feature to memory resource controller. When you
714 enable this, you can limit mem+swap usage per cgroup. In other words,
715 when you disable this, memory resource controller has no cares to
716 usage of swap...a process can exhaust all of the swap. This extension
717 is useful when you want to avoid exhaustion swap but this itself
718 adds more overheads and consumes memory for remembering information.
719 Especially if you use 32bit system or small memory system, please
720 be careful about enabling this. When memory resource controller
721 is disabled by boot option, this will be automatically disabled and
722 there will be no overhead from this. Even when you set this config=y,
723 if boot option "swapaccount=0" is set, swap will not be accounted.
724 Now, memory usage of swap_cgroup is 2 bytes per entry. If swap page
725 size is 4096bytes, 512k per 1Gbytes of swap.
726 config CGROUP_MEM_RES_CTLR_SWAP_ENABLED
727 bool "Memory Resource Controller Swap Extension enabled by default"
728 depends on CGROUP_MEM_RES_CTLR_SWAP
731 Memory Resource Controller Swap Extension comes with its price in
732 a bigger memory consumption. General purpose distribution kernels
733 which want to enable the feature but keep it disabled by default
734 and let the user enable it by swapaccount boot command line
735 parameter should have this option unselected.
736 For those who want to have the feature enabled by default should
737 select this option (if, for some reason, they need to disable it
738 then swapaccount=0 does the trick).
739 config CGROUP_MEM_RES_CTLR_KMEM
740 bool "Memory Resource Controller Kernel Memory accounting (EXPERIMENTAL)"
741 depends on CGROUP_MEM_RES_CTLR && EXPERIMENTAL
744 The Kernel Memory extension for Memory Resource Controller can limit
745 the amount of memory used by kernel objects in the system. Those are
746 fundamentally different from the entities handled by the standard
747 Memory Controller, which are page-based, and can be swapped. Users of
748 the kmem extension can use it to guarantee that no group of processes
749 will ever exhaust kernel resources alone.
752 bool "Enable perf_event per-cpu per-container group (cgroup) monitoring"
753 depends on PERF_EVENTS && CGROUPS
755 This option extends the per-cpu mode to restrict monitoring to
756 threads which belong to the cgroup specified and run on the
761 menuconfig CGROUP_SCHED
762 bool "Group CPU scheduler"
765 This feature lets CPU scheduler recognize task groups and control CPU
766 bandwidth allocation to such task groups. It uses cgroups to group
770 config FAIR_GROUP_SCHED
771 bool "Group scheduling for SCHED_OTHER"
772 depends on CGROUP_SCHED
776 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
777 depends on EXPERIMENTAL
778 depends on FAIR_GROUP_SCHED
781 This option allows users to define CPU bandwidth rates (limits) for
782 tasks running within the fair group scheduler. Groups with no limit
783 set are considered to be unconstrained and will run with no
785 See tip/Documentation/scheduler/sched-bwc.txt for more information.
787 config RT_GROUP_SCHED
788 bool "Group scheduling for SCHED_RR/FIFO"
789 depends on EXPERIMENTAL
790 depends on CGROUP_SCHED
793 This feature lets you explicitly allocate real CPU bandwidth
794 to task groups. If enabled, it will also make it impossible to
795 schedule realtime tasks for non-root users until you allocate
796 realtime bandwidth for them.
797 See Documentation/scheduler/sched-rt-group.txt for more information.
802 tristate "Block IO controller"
806 Generic block IO controller cgroup interface. This is the common
807 cgroup interface which should be used by various IO controlling
810 Currently, CFQ IO scheduler uses it to recognize task groups and
811 control disk bandwidth allocation (proportional time slice allocation)
812 to such task groups. It is also used by bio throttling logic in
813 block layer to implement upper limit in IO rates on a device.
815 This option only enables generic Block IO controller infrastructure.
816 One needs to also enable actual IO controlling logic/policy. For
817 enabling proportional weight division of disk bandwidth in CFQ, set
818 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
819 CONFIG_BLK_DEV_THROTTLING=y.
821 See Documentation/cgroups/blkio-controller.txt for more information.
823 config DEBUG_BLK_CGROUP
824 bool "Enable Block IO controller debugging"
825 depends on BLK_CGROUP
828 Enable some debugging help. Currently it exports additional stat
829 files in a cgroup which can be useful for debugging.
833 config CHECKPOINT_RESTORE
834 bool "Checkpoint/restore support" if EXPERT
837 Enables additional kernel features in a sake of checkpoint/restore.
838 In particular it adds auxiliary prctl codes to setup process text,
839 data and heap segment sizes, and a few additional /proc filesystem
842 If unsure, say N here.
844 menuconfig NAMESPACES
845 bool "Namespaces support" if EXPERT
848 Provides the way to make tasks work with different objects using
849 the same id. For example same IPC id may refer to different objects
850 or same user id or pid may refer to different tasks when used in
851 different namespaces.
859 In this namespace tasks see different info provided with the
864 depends on (SYSVIPC || POSIX_MQUEUE)
867 In this namespace tasks work with IPC ids which correspond to
868 different IPC objects in different namespaces.
871 bool "User namespace (EXPERIMENTAL)"
872 depends on EXPERIMENTAL
875 This allows containers, i.e. vservers, to use user namespaces
876 to provide different user info for different servers.
880 bool "PID Namespaces"
883 Support process id namespaces. This allows having multiple
884 processes with the same pid as long as they are in different
885 pid namespaces. This is a building block of containers.
888 bool "Network namespace"
892 Allow user space to create what appear to be multiple instances
893 of the network stack.
897 config SCHED_AUTOGROUP
898 bool "Automatic process group scheduling"
902 select FAIR_GROUP_SCHED
904 This option optimizes the scheduler for common desktop workloads by
905 automatically creating and populating task groups. This separation
906 of workloads isolates aggressive CPU burners (like build jobs) from
907 desktop applications. Task group autogeneration is currently based
913 config SYSFS_DEPRECATED
914 bool "Enable deprecated sysfs features to support old userspace tools"
918 This option adds code that switches the layout of the "block" class
919 devices, to not show up in /sys/class/block/, but only in
922 This switch is only active when the sysfs.deprecated=1 boot option is
923 passed or the SYSFS_DEPRECATED_V2 option is set.
925 This option allows new kernels to run on old distributions and tools,
926 which might get confused by /sys/class/block/. Since 2007/2008 all
927 major distributions and tools handle this just fine.
929 Recent distributions and userspace tools after 2009/2010 depend on
930 the existence of /sys/class/block/, and will not work with this
933 Only if you are using a new kernel on an old distribution, you might
936 config SYSFS_DEPRECATED_V2
937 bool "Enable deprecated sysfs features by default"
940 depends on SYSFS_DEPRECATED
942 Enable deprecated sysfs by default.
944 See the CONFIG_SYSFS_DEPRECATED option for more details about this
947 Only if you are using a new kernel on an old distribution, you might
948 need to say Y here. Even then, odds are you would not need it
949 enabled, you can always pass the boot option if absolutely necessary.
952 bool "Kernel->user space relay support (formerly relayfs)"
954 This option enables support for relay interface support in
955 certain file systems (such as debugfs).
956 It is designed to provide an efficient mechanism for tools and
957 facilities to relay large amounts of data from kernel space to
962 config BLK_DEV_INITRD
963 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
964 depends on BROKEN || !FRV
966 The initial RAM filesystem is a ramfs which is loaded by the
967 boot loader (loadlin or lilo) and that is mounted as root
968 before the normal boot procedure. It is typically used to
969 load modules needed to mount the "real" root file system,
970 etc. See <file:Documentation/initrd.txt> for details.
972 If RAM disk support (BLK_DEV_RAM) is also included, this
973 also enables initial RAM disk (initrd) support and adds
974 15 Kbytes (more on some other architectures) to the kernel size.
984 config CC_OPTIMIZE_FOR_SIZE
985 bool "Optimize for size"
987 Enabling this option will pass "-Os" instead of "-O2" to gcc
988 resulting in a smaller kernel.
999 bool "Configure standard kernel features (expert users)"
1000 # Unhide debug options, to make the on-by-default options visible
1003 This option allows certain base kernel options and settings
1004 to be disabled or tweaked. This is for specialized
1005 environments which can tolerate a "non-standard" kernel.
1006 Only use this if you really know what you are doing.
1009 bool "Enable 16-bit UID system calls" if EXPERT
1010 depends on ARM || BLACKFIN || CRIS || FRV || H8300 || X86_32 || M68K || (S390 && !64BIT) || SUPERH || SPARC32 || (SPARC64 && COMPAT) || UML || (X86_64 && IA32_EMULATION)
1013 This enables the legacy 16-bit UID syscall wrappers.
1015 config SYSCTL_SYSCALL
1016 bool "Sysctl syscall support" if EXPERT
1017 depends on PROC_SYSCTL
1021 sys_sysctl uses binary paths that have been found challenging
1022 to properly maintain and use. The interface in /proc/sys
1023 using paths with ascii names is now the primary path to this
1026 Almost nothing using the binary sysctl interface so if you are
1027 trying to save some space it is probably safe to disable this,
1028 making your kernel marginally smaller.
1030 If unsure say N here.
1033 bool "Load all symbols for debugging/ksymoops" if EXPERT
1036 Say Y here to let the kernel print out symbolic crash information and
1037 symbolic stack backtraces. This increases the size of the kernel
1038 somewhat, as all symbols have to be loaded into the kernel image.
1041 bool "Include all symbols in kallsyms"
1042 depends on DEBUG_KERNEL && KALLSYMS
1044 Normally kallsyms only contains the symbols of functions for nicer
1045 OOPS messages and backtraces (i.e., symbols from the text and inittext
1046 sections). This is sufficient for most cases. And only in very rare
1047 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1048 names of variables from the data sections, etc).
1050 This option makes sure that all symbols are loaded into the kernel
1051 image (i.e., symbols from all sections) in cost of increased kernel
1052 size (depending on the kernel configuration, it may be 300KiB or
1053 something like this).
1055 Say N unless you really need all symbols.
1058 bool "Support for hot-pluggable devices" if EXPERT
1061 This option is provided for the case where no hotplug or uevent
1062 capabilities is wanted by the kernel. You should only consider
1063 disabling this option for embedded systems that do not use modules, a
1064 dynamic /dev tree, or dynamic device discovery. Just say Y.
1068 bool "Enable support for printk" if EXPERT
1070 This option enables normal printk support. Removing it
1071 eliminates most of the message strings from the kernel image
1072 and makes the kernel more or less silent. As this makes it
1073 very difficult to diagnose system problems, saying N here is
1074 strongly discouraged.
1077 bool "BUG() support" if EXPERT
1080 Disabling this option eliminates support for BUG and WARN, reducing
1081 the size of your kernel image and potentially quietly ignoring
1082 numerous fatal conditions. You should only consider disabling this
1083 option for embedded systems with no facilities for reporting errors.
1088 bool "Enable ELF core dumps" if EXPERT
1090 Enable support for generating core dumps. Disabling saves about 4k.
1093 config PCSPKR_PLATFORM
1094 bool "Enable PC-Speaker support" if EXPERT
1095 depends on HAVE_PCSPKR_PLATFORM
1099 This option allows to disable the internal PC-Speaker
1100 support, saving some memory.
1102 config HAVE_PCSPKR_PLATFORM
1107 bool "Enable full-sized data structures for core" if EXPERT
1109 Disabling this option reduces the size of miscellaneous core
1110 kernel data structures. This saves memory on small machines,
1111 but may reduce performance.
1114 bool "Enable futex support" if EXPERT
1118 Disabling this option will cause the kernel to be built without
1119 support for "fast userspace mutexes". The resulting kernel may not
1120 run glibc-based applications correctly.
1123 bool "Enable eventpoll support" if EXPERT
1127 Disabling this option will cause the kernel to be built without
1128 support for epoll family of system calls.
1131 bool "Enable signalfd() system call" if EXPERT
1135 Enable the signalfd() system call that allows to receive signals
1136 on a file descriptor.
1141 bool "Enable timerfd() system call" if EXPERT
1145 Enable the timerfd() system call that allows to receive timer
1146 events on a file descriptor.
1151 bool "Enable eventfd() system call" if EXPERT
1155 Enable the eventfd() system call that allows to receive both
1156 kernel notification (ie. KAIO) or userspace notifications.
1161 bool "Use full shmem filesystem" if EXPERT
1165 The shmem is an internal filesystem used to manage shared memory.
1166 It is backed by swap and manages resource limits. It is also exported
1167 to userspace as tmpfs if TMPFS is enabled. Disabling this
1168 option replaces shmem and tmpfs with the much simpler ramfs code,
1169 which may be appropriate on small systems without swap.
1172 bool "Enable AIO support" if EXPERT
1175 This option enables POSIX asynchronous I/O which may by used
1176 by some high performance threaded applications. Disabling
1177 this option saves about 7k.
1180 bool "Embedded system"
1183 This option should be enabled if compiling the kernel for
1184 an embedded system so certain expert options are available
1187 config HAVE_PERF_EVENTS
1190 See tools/perf/design.txt for details.
1192 config PERF_USE_VMALLOC
1195 See tools/perf/design.txt for details
1197 menu "Kernel Performance Events And Counters"
1200 bool "Kernel performance events and counters"
1201 default y if (PROFILING || PERF_COUNTERS)
1202 depends on HAVE_PERF_EVENTS
1206 Enable kernel support for various performance events provided
1207 by software and hardware.
1209 Software events are supported either built-in or via the
1210 use of generic tracepoints.
1212 Most modern CPUs support performance events via performance
1213 counter registers. These registers count the number of certain
1214 types of hw events: such as instructions executed, cachemisses
1215 suffered, or branches mis-predicted - without slowing down the
1216 kernel or applications. These registers can also trigger interrupts
1217 when a threshold number of events have passed - and can thus be
1218 used to profile the code that runs on that CPU.
1220 The Linux Performance Event subsystem provides an abstraction of
1221 these software and hardware event capabilities, available via a
1222 system call and used by the "perf" utility in tools/perf/. It
1223 provides per task and per CPU counters, and it provides event
1224 capabilities on top of those.
1228 config PERF_COUNTERS
1229 bool "Kernel performance counters (old config option)"
1230 depends on HAVE_PERF_EVENTS
1232 This config has been obsoleted by the PERF_EVENTS
1233 config option - please see that one for details.
1235 It has no effect on the kernel whether you enable
1236 it or not, it is a compatibility placeholder.
1240 config DEBUG_PERF_USE_VMALLOC
1242 bool "Debug: use vmalloc to back perf mmap() buffers"
1243 depends on PERF_EVENTS && DEBUG_KERNEL
1244 select PERF_USE_VMALLOC
1246 Use vmalloc memory to back perf mmap() buffers.
1248 Mostly useful for debugging the vmalloc code on platforms
1249 that don't require it.
1255 config VM_EVENT_COUNTERS
1257 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1259 VM event counters are needed for event counts to be shown.
1260 This option allows the disabling of the VM event counters
1261 on EXPERT systems. /proc/vmstat will only show page counts
1262 if VM event counters are disabled.
1266 bool "Enable PCI quirk workarounds" if EXPERT
1269 This enables workarounds for various PCI chipset
1270 bugs/quirks. Disable this only if your target machine is
1271 unaffected by PCI quirks.
1275 bool "Enable SLUB debugging support" if EXPERT
1276 depends on SLUB && SYSFS
1278 SLUB has extensive debug support features. Disabling these can
1279 result in significant savings in code size. This also disables
1280 SLUB sysfs support. /sys/slab will not exist and there will be
1281 no support for cache validation etc.
1284 bool "Disable heap randomization"
1287 Randomizing heap placement makes heap exploits harder, but it
1288 also breaks ancient binaries (including anything libc5 based).
1289 This option changes the bootup default to heap randomization
1290 disabled, and can be overridden at runtime by setting
1291 /proc/sys/kernel/randomize_va_space to 2.
1293 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1296 prompt "Choose SLAB allocator"
1299 This option allows to select a slab allocator.
1304 The regular slab allocator that is established and known to work
1305 well in all environments. It organizes cache hot objects in
1306 per cpu and per node queues.
1309 bool "SLUB (Unqueued Allocator)"
1311 SLUB is a slab allocator that minimizes cache line usage
1312 instead of managing queues of cached objects (SLAB approach).
1313 Per cpu caching is realized using slabs of objects instead
1314 of queues of objects. SLUB can use memory efficiently
1315 and has enhanced diagnostics. SLUB is the default choice for
1320 bool "SLOB (Simple Allocator)"
1322 SLOB replaces the stock allocator with a drastically simpler
1323 allocator. SLOB is generally more space efficient but
1324 does not perform as well on large systems.
1328 config MMAP_ALLOW_UNINITIALIZED
1329 bool "Allow mmapped anonymous memory to be uninitialized"
1330 depends on EXPERT && !MMU
1333 Normally, and according to the Linux spec, anonymous memory obtained
1334 from mmap() has it's contents cleared before it is passed to
1335 userspace. Enabling this config option allows you to request that
1336 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1337 providing a huge performance boost. If this option is not enabled,
1338 then the flag will be ignored.
1340 This is taken advantage of by uClibc's malloc(), and also by
1341 ELF-FDPIC binfmt's brk and stack allocator.
1343 Because of the obvious security issues, this option should only be
1344 enabled on embedded devices where you control what is run in
1345 userspace. Since that isn't generally a problem on no-MMU systems,
1346 it is normally safe to say Y here.
1348 See Documentation/nommu-mmap.txt for more information.
1351 bool "Profiling support"
1353 Say Y here to enable the extended profiling support mechanisms used
1354 by profilers such as OProfile.
1357 # Place an empty function call at each tracepoint site. Can be
1358 # dynamically changed for a probe function.
1363 source "arch/Kconfig"
1365 endmenu # General setup
1367 config HAVE_GENERIC_DMA_COHERENT
1374 depends on SLAB || SLUB_DEBUG
1382 default 0 if BASE_FULL
1383 default 1 if !BASE_FULL
1386 bool "Enable loadable module support"
1388 Kernel modules are small pieces of compiled code which can
1389 be inserted in the running kernel, rather than being
1390 permanently built into the kernel. You use the "modprobe"
1391 tool to add (and sometimes remove) them. If you say Y here,
1392 many parts of the kernel can be built as modules (by
1393 answering M instead of Y where indicated): this is most
1394 useful for infrequently used options which are not required
1395 for booting. For more information, see the man pages for
1396 modprobe, lsmod, modinfo, insmod and rmmod.
1398 If you say Y here, you will need to run "make
1399 modules_install" to put the modules under /lib/modules/
1400 where modprobe can find them (you may need to be root to do
1407 config MODULE_FORCE_LOAD
1408 bool "Forced module loading"
1411 Allow loading of modules without version information (ie. modprobe
1412 --force). Forced module loading sets the 'F' (forced) taint flag and
1413 is usually a really bad idea.
1415 config MODULE_UNLOAD
1416 bool "Module unloading"
1418 Without this option you will not be able to unload any
1419 modules (note that some modules may not be unloadable
1420 anyway), which makes your kernel smaller, faster
1421 and simpler. If unsure, say Y.
1423 config MODULE_FORCE_UNLOAD
1424 bool "Forced module unloading"
1425 depends on MODULE_UNLOAD && EXPERIMENTAL
1427 This option allows you to force a module to unload, even if the
1428 kernel believes it is unsafe: the kernel will remove the module
1429 without waiting for anyone to stop using it (using the -f option to
1430 rmmod). This is mainly for kernel developers and desperate users.
1434 bool "Module versioning support"
1436 Usually, you have to use modules compiled with your kernel.
1437 Saying Y here makes it sometimes possible to use modules
1438 compiled for different kernels, by adding enough information
1439 to the modules to (hopefully) spot any changes which would
1440 make them incompatible with the kernel you are running. If
1443 config MODULE_SRCVERSION_ALL
1444 bool "Source checksum for all modules"
1446 Modules which contain a MODULE_VERSION get an extra "srcversion"
1447 field inserted into their modinfo section, which contains a
1448 sum of the source files which made it. This helps maintainers
1449 see exactly which source was used to build a module (since
1450 others sometimes change the module source without updating
1451 the version). With this option, such a "srcversion" field
1452 will be created for all modules. If unsure, say N.
1456 config INIT_ALL_POSSIBLE
1459 Back when each arch used to define their own cpu_online_mask and
1460 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
1461 with all 1s, and others with all 0s. When they were centralised,
1462 it was better to provide this option than to break all the archs
1463 and have several arch maintainers pursuing me down dark alleys.
1468 depends on (SMP && MODULE_UNLOAD) || HOTPLUG_CPU
1470 Need stop_machine() primitive.
1472 source "block/Kconfig"
1474 config PREEMPT_NOTIFIERS
1481 source "kernel/Kconfig.locks"