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 prompt "CPU Scheduler"
36 Select the CPU Scheduler to be used.
39 bool "Brain Fuck Scheduler (BFS)"
41 The Brain Fuck CPU Scheduler for excellent interactivity and
42 responsiveness on the desktop and solid scalability on normal
43 hardware. Not recommended for 4096 CPUs.
45 Currently incompatible with the Group CPU scheduler, and RCU TORTURE
46 TEST so these options are disabled.
51 bool "Completely Fair Scheduler (CFS)"
53 This is the default cpu scheduler found in the upstream kernel.
56 config SCHED_BFS_AUTOISO
57 bool "Automatically use SCHED_ISO policy for X"
61 Selecting this option will automatically use the SCHED_ISO scheduling
62 policy for X, resulting in an interactivity boost. This *may* cause
63 things like skipping sound on audio applications that are not run
66 Tasks (including X) can be run as sched_iso manually using schedtool.
70 prompt "Zen-Tune Profile"
73 Select the tunable profile to be used.
78 This option sets tunables to default.
81 BACKGROUND_DIRTY_RATIO: 10
87 SCHED LATENCY GRANULARITY: 0.75 ms
93 This option sets tunables to optimal desktop values.
94 Changes to these values can be adjusted in include/linux/zentune.h
97 BACKGROUND_DIRTY_RATIO: 20
103 SCHED LATENCY GRANULARITY: 0.3 ms
109 bool "Prompt for development and/or incomplete code/drivers"
111 Some of the various things that Linux supports (such as network
112 drivers, file systems, network protocols, etc.) can be in a state
113 of development where the functionality, stability, or the level of
114 testing is not yet high enough for general use. This is usually
115 known as the "alpha-test" phase among developers. If a feature is
116 currently in alpha-test, then the developers usually discourage
117 uninformed widespread use of this feature by the general public to
118 avoid "Why doesn't this work?" type mail messages. However, active
119 testing and use of these systems is welcomed. Just be aware that it
120 may not meet the normal level of reliability or it may fail to work
121 in some special cases. Detailed bug reports from people familiar
122 with the kernel internals are usually welcomed by the developers
123 (before submitting bug reports, please read the documents
124 <file:README>, <file:MAINTAINERS>, <file:REPORTING-BUGS>,
125 <file:Documentation/BUG-HUNTING>, and
126 <file:Documentation/oops-tracing.txt> in the kernel source).
128 This option will also make obsoleted drivers available. These are
129 drivers that have been replaced by something else, and/or are
130 scheduled to be removed in a future kernel release.
132 Unless you intend to help test and develop a feature or driver that
133 falls into this category, or you have a situation that requires
134 using these features, you should probably say N here, which will
135 cause the configurator to present you with fewer choices. If
136 you say Y here, you will be offered the choice of using features or
137 drivers that are currently considered to be in the alpha-test phase.
144 depends on BROKEN || !SMP
147 config INIT_ENV_ARG_LIMIT
152 Maximum of each of the number of arguments and environment
153 variables passed to init from the kernel command line.
157 string "Cross-compiler tool prefix"
159 Same as running 'make CROSS_COMPILE=prefix-' but stored for
160 default make runs in this kernel build directory. You don't
161 need to set this unless you want the configured kernel build
162 directory to select the cross-compiler automatically.
165 string "Local version - append to kernel release"
167 Append an extra string to the end of your kernel version.
168 This will show up when you type uname, for example.
169 The string you set here will be appended after the contents of
170 any files with a filename matching localversion* in your
171 object and source tree, in that order. Your total string can
172 be a maximum of 64 characters.
174 config LOCALVERSION_AUTO
175 bool "Automatically append version information to the version string"
178 This will try to automatically determine if the current tree is a
179 release tree by looking for git tags that belong to the current
180 top of tree revision.
182 A string of the format -gxxxxxxxx will be added to the localversion
183 if a git-based tree is found. The string generated by this will be
184 appended after any matching localversion* files, and after the value
185 set in CONFIG_LOCALVERSION.
187 (The actual string used here is the first eight characters produced
188 by running the command:
190 $ git rev-parse --verify HEAD
192 which is done within the script "scripts/setlocalversion".)
194 config HAVE_KERNEL_GZIP
197 config HAVE_KERNEL_BZIP2
200 config HAVE_KERNEL_LZMA
203 config HAVE_KERNEL_XZ
206 config HAVE_KERNEL_LZO
210 prompt "Kernel compression mode"
212 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO
214 The linux kernel is a kind of self-extracting executable.
215 Several compression algorithms are available, which differ
216 in efficiency, compression and decompression speed.
217 Compression speed is only relevant when building a kernel.
218 Decompression speed is relevant at each boot.
220 If you have any problems with bzip2 or lzma compressed
221 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
222 version of this functionality (bzip2 only), for 2.4, was
223 supplied by Christian Ludwig)
225 High compression options are mostly useful for users, who
226 are low on disk space (embedded systems), but for whom ram
229 If in doubt, select 'gzip'
233 depends on HAVE_KERNEL_GZIP
235 The old and tried gzip compression. It provides a good balance
236 between compression ratio and decompression speed.
240 depends on HAVE_KERNEL_BZIP2
242 Its compression ratio and speed is intermediate.
243 Decompression speed is slowest among the three. The kernel
244 size is about 10% smaller with bzip2, in comparison to gzip.
245 Bzip2 uses a large amount of memory. For modern kernels you
246 will need at least 8MB RAM or more for booting.
250 depends on HAVE_KERNEL_LZMA
252 The most recent compression algorithm.
253 Its ratio is best, decompression speed is between the other
254 two. Compression is slowest. The kernel size is about 33%
255 smaller with LZMA in comparison to gzip.
259 depends on HAVE_KERNEL_XZ
261 XZ uses the LZMA2 algorithm and instruction set specific
262 BCJ filters which can improve compression ratio of executable
263 code. The size of the kernel is about 30% smaller with XZ in
264 comparison to gzip. On architectures for which there is a BCJ
265 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
266 will create a few percent smaller kernel than plain LZMA.
268 The speed is about the same as with LZMA: The decompression
269 speed of XZ is better than that of bzip2 but worse than gzip
270 and LZO. Compression is slow.
274 depends on HAVE_KERNEL_LZO
276 Its compression ratio is the poorest among the 4. The kernel
277 size is about 10% bigger than gzip; however its speed
278 (both compression and decompression) is the fastest.
282 config DEFAULT_HOSTNAME
283 string "Default hostname"
286 This option determines the default system hostname before userspace
287 calls sethostname(2). The kernel traditionally uses "(none)" here,
288 but you may wish to use a different default here to make a minimal
289 system more usable with less configuration.
292 bool "Support for paging of anonymous memory (swap)"
293 depends on MMU && BLOCK
296 This option allows you to choose whether you want to have support
297 for so called swap devices or swap files in your kernel that are
298 used to provide more virtual memory than the actual RAM present
299 in your computer. If unsure say Y.
304 Inter Process Communication is a suite of library functions and
305 system calls which let processes (running programs) synchronize and
306 exchange information. It is generally considered to be a good thing,
307 and some programs won't run unless you say Y here. In particular, if
308 you want to run the DOS emulator dosemu under Linux (read the
309 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
310 you'll need to say Y here.
312 You can find documentation about IPC with "info ipc" and also in
313 section 6.4 of the Linux Programmer's Guide, available from
314 <http://www.tldp.org/guides.html>.
316 config SYSVIPC_SYSCTL
323 bool "POSIX Message Queues"
324 depends on NET && EXPERIMENTAL
326 POSIX variant of message queues is a part of IPC. In POSIX message
327 queues every message has a priority which decides about succession
328 of receiving it by a process. If you want to compile and run
329 programs written e.g. for Solaris with use of its POSIX message
330 queues (functions mq_*) say Y here.
332 POSIX message queues are visible as a filesystem called 'mqueue'
333 and can be mounted somewhere if you want to do filesystem
334 operations on message queues.
338 config POSIX_MQUEUE_SYSCTL
340 depends on POSIX_MQUEUE
344 config BSD_PROCESS_ACCT
345 bool "BSD Process Accounting"
347 If you say Y here, a user level program will be able to instruct the
348 kernel (via a special system call) to write process accounting
349 information to a file: whenever a process exits, information about
350 that process will be appended to the file by the kernel. The
351 information includes things such as creation time, owning user,
352 command name, memory usage, controlling terminal etc. (the complete
353 list is in the struct acct in <file:include/linux/acct.h>). It is
354 up to the user level program to do useful things with this
355 information. This is generally a good idea, so say Y.
357 config BSD_PROCESS_ACCT_V3
358 bool "BSD Process Accounting version 3 file format"
359 depends on BSD_PROCESS_ACCT
362 If you say Y here, the process accounting information is written
363 in a new file format that also logs the process IDs of each
364 process and it's parent. Note that this file format is incompatible
365 with previous v0/v1/v2 file formats, so you will need updated tools
366 for processing it. A preliminary version of these tools is available
367 at <http://www.gnu.org/software/acct/>.
370 bool "open by fhandle syscalls"
373 If you say Y here, a user level program will be able to map
374 file names to handle and then later use the handle for
375 different file system operations. This is useful in implementing
376 userspace file servers, which now track files using handles instead
377 of names. The handle would remain the same even if file names
378 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
382 bool "Export task/process statistics through netlink (EXPERIMENTAL)"
386 Export selected statistics for tasks/processes through the
387 generic netlink interface. Unlike BSD process accounting, the
388 statistics are available during the lifetime of tasks/processes as
389 responses to commands. Like BSD accounting, they are sent to user
394 config TASK_DELAY_ACCT
395 bool "Enable per-task delay accounting (EXPERIMENTAL)"
398 Collect information on time spent by a task waiting for system
399 resources like cpu, synchronous block I/O completion and swapping
400 in pages. Such statistics can help in setting a task's priorities
401 relative to other tasks for cpu, io, rss limits etc.
406 bool "Enable extended accounting over taskstats (EXPERIMENTAL)"
409 Collect extended task accounting data and send the data
410 to userland for processing over the taskstats interface.
414 config TASK_IO_ACCOUNTING
415 bool "Enable per-task storage I/O accounting (EXPERIMENTAL)"
416 depends on TASK_XACCT
418 Collect information on the number of bytes of storage I/O which this
424 bool "Auditing support"
427 Enable auditing infrastructure that can be used with another
428 kernel subsystem, such as SELinux (which requires this for
429 logging of avc messages output). Does not do system-call
430 auditing without CONFIG_AUDITSYSCALL.
433 bool "Enable system-call auditing support"
434 depends on AUDIT && (X86 || PPC || S390 || IA64 || UML || SPARC64 || SUPERH || ARM)
435 default y if SECURITY_SELINUX
437 Enable low-overhead system-call auditing infrastructure that
438 can be used independently or with another kernel subsystem,
443 depends on AUDITSYSCALL
448 depends on AUDITSYSCALL
451 config AUDIT_LOGINUID_IMMUTABLE
452 bool "Make audit loginuid immutable"
455 The config option toggles if a task setting its loginuid requires
456 CAP_SYS_AUDITCONTROL or if that task should require no special permissions
457 but should instead only allow setting its loginuid if it was never
458 previously set. On systems which use systemd or a similar central
459 process to restart login services this should be set to true. On older
460 systems in which an admin would typically have to directly stop and
461 start processes this should be set to false. Setting this to true allows
462 one to drop potentially dangerous capabilites from the login tasks,
463 but may not be backwards compatible with older init systems.
465 source "kernel/irq/Kconfig"
470 prompt "RCU Implementation"
474 bool "Tree-based hierarchical RCU"
475 depends on !PREEMPT && SMP
477 This option selects the RCU implementation that is
478 designed for very large SMP system with hundreds or
479 thousands of CPUs. It also scales down nicely to
482 config TREE_PREEMPT_RCU
483 bool "Preemptible tree-based hierarchical RCU"
484 depends on PREEMPT && SMP
486 This option selects the RCU implementation that is
487 designed for very large SMP systems with hundreds or
488 thousands of CPUs, but for which real-time response
489 is also required. It also scales down nicely to
493 bool "UP-only small-memory-footprint RCU"
494 depends on !PREEMPT && !SMP
496 This option selects the RCU implementation that is
497 designed for UP systems from which real-time response
498 is not required. This option greatly reduces the
499 memory footprint of RCU.
501 config TINY_PREEMPT_RCU
502 bool "Preemptible UP-only small-memory-footprint RCU"
503 depends on PREEMPT && !SMP
505 This option selects the RCU implementation that is designed
506 for real-time UP systems. This option greatly reduces the
507 memory footprint of RCU.
512 def_bool ( TREE_PREEMPT_RCU || TINY_PREEMPT_RCU )
514 This option enables preemptible-RCU code that is common between
515 the TREE_PREEMPT_RCU and TINY_PREEMPT_RCU implementations.
518 bool "Enable tracing for RCU"
520 This option provides tracing in RCU which presents stats
521 in debugfs for debugging RCU implementation.
523 Say Y here if you want to enable RCU tracing
524 Say N if you are unsure.
527 int "Tree-based hierarchical RCU fanout value"
530 depends on TREE_RCU || TREE_PREEMPT_RCU
534 This option controls the fanout of hierarchical implementations
535 of RCU, allowing RCU to work efficiently on machines with
536 large numbers of CPUs. This value must be at least the fourth
537 root of NR_CPUS, which allows NR_CPUS to be insanely large.
538 The default value of RCU_FANOUT should be used for production
539 systems, but if you are stress-testing the RCU implementation
540 itself, small RCU_FANOUT values allow you to test large-system
541 code paths on small(er) systems.
543 Select a specific number if testing RCU itself.
544 Take the default if unsure.
546 config RCU_FANOUT_EXACT
547 bool "Disable tree-based hierarchical RCU auto-balancing"
548 depends on TREE_RCU || TREE_PREEMPT_RCU
551 This option forces use of the exact RCU_FANOUT value specified,
552 regardless of imbalances in the hierarchy. This is useful for
553 testing RCU itself, and might one day be useful on systems with
554 strong NUMA behavior.
556 Without RCU_FANOUT_EXACT, the code will balance the hierarchy.
560 config RCU_FAST_NO_HZ
561 bool "Accelerate last non-dyntick-idle CPU's grace periods"
562 depends on NO_HZ && SMP
565 This option causes RCU to attempt to accelerate grace periods
566 in order to allow CPUs to enter dynticks-idle state more
567 quickly. On the other hand, this option increases the overhead
568 of the dynticks-idle checking, particularly on systems with
569 large numbers of CPUs.
571 Say Y if energy efficiency is critically important, particularly
572 if you have relatively few CPUs.
574 Say N if you are unsure.
576 config TREE_RCU_TRACE
577 def_bool RCU_TRACE && ( TREE_RCU || TREE_PREEMPT_RCU )
580 This option provides tracing for the TREE_RCU and
581 TREE_PREEMPT_RCU implementations, permitting Makefile to
582 trivially select kernel/rcutree_trace.c.
585 bool "Enable RCU priority boosting"
586 depends on RT_MUTEXES && PREEMPT_RCU
589 This option boosts the priority of preempted RCU readers that
590 block the current preemptible RCU grace period for too long.
591 This option also prevents heavy loads from blocking RCU
592 callback invocation for all flavors of RCU.
594 Say Y here if you are working with real-time apps or heavy loads
595 Say N here if you are unsure.
597 config RCU_BOOST_PRIO
598 int "Real-time priority to boost RCU readers to"
603 This option specifies the real-time priority to which preempted
604 RCU readers are to be boosted. If you are working with CPU-bound
605 real-time applications, you should specify a priority higher then
606 the highest-priority CPU-bound application.
608 Specify the real-time priority, or take the default if unsure.
610 config RCU_BOOST_DELAY
611 int "Milliseconds to delay boosting after RCU grace-period start"
616 This option specifies the time to wait after the beginning of
617 a given grace period before priority-boosting preempted RCU
618 readers blocking that grace period. Note that any RCU reader
619 blocking an expedited RCU grace period is boosted immediately.
621 Accept the default if unsure.
623 endmenu # "RCU Subsystem"
626 tristate "Kernel .config support"
628 This option enables the complete Linux kernel ".config" file
629 contents to be saved in the kernel. It provides documentation
630 of which kernel options are used in a running kernel or in an
631 on-disk kernel. This information can be extracted from the kernel
632 image file with the script scripts/extract-ikconfig and used as
633 input to rebuild the current kernel or to build another kernel.
634 It can also be extracted from a running kernel by reading
635 /proc/config.gz if enabled (below).
638 bool "Enable access to .config through /proc/config.gz"
639 depends on IKCONFIG && PROC_FS
641 This option enables access to the kernel configuration file
642 through /proc/config.gz.
645 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
649 Select kernel log buffer size as a power of 2.
659 # Architectures with an unreliable sched_clock() should select this:
661 config HAVE_UNSTABLE_SCHED_CLOCK
665 boolean "Control Group support"
668 This option adds support for grouping sets of processes together, for
669 use with process control subsystems such as Cpusets, CFS, memory
670 controls or device isolation.
672 - Documentation/scheduler/sched-design-CFS.txt (CFS)
673 - Documentation/cgroups/ (features for grouping, isolation
674 and resource control)
681 bool "Example debug cgroup subsystem"
684 This option enables a simple cgroup subsystem that
685 exports useful debugging information about the cgroups
690 config CGROUP_FREEZER
691 bool "Freezer cgroup subsystem"
693 Provides a way to freeze and unfreeze all tasks in a
697 bool "Device controller for cgroups"
699 Provides a cgroup implementing whitelists for devices which
700 a process in the cgroup can mknod or open.
703 bool "Cpuset support"
705 This option will let you create and manage CPUSETs which
706 allow dynamically partitioning a system into sets of CPUs and
707 Memory Nodes and assigning tasks to run only within those sets.
708 This is primarily useful on large SMP or NUMA systems.
712 config PROC_PID_CPUSET
713 bool "Include legacy /proc/<pid>/cpuset file"
717 config CGROUP_CPUACCT
718 bool "Simple CPU accounting cgroup subsystem"
719 depends on !SCHED_BFS
721 Provides a simple Resource Controller for monitoring the
722 total CPU consumed by the tasks in a cgroup.
724 config RESOURCE_COUNTERS
725 bool "Resource counters"
727 This option enables controller independent resource accounting
728 infrastructure that works with cgroups.
730 config CGROUP_MEM_RES_CTLR
731 bool "Memory Resource Controller for Control Groups"
732 depends on RESOURCE_COUNTERS
735 Provides a memory resource controller that manages both anonymous
736 memory and page cache. (See Documentation/cgroups/memory.txt)
738 Note that setting this option increases fixed memory overhead
739 associated with each page of memory in the system. By this,
740 20(40)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory
741 usage tracking struct at boot. Total amount of this is printed out
744 Only enable when you're ok with these trade offs and really
745 sure you need the memory resource controller. Even when you enable
746 this, you can set "cgroup_disable=memory" at your boot option to
747 disable memory resource controller and you can avoid overheads.
748 (and lose benefits of memory resource controller)
750 This config option also selects MM_OWNER config option, which
751 could in turn add some fork/exit overhead.
753 config CGROUP_MEM_RES_CTLR_DISABLED
754 bool "Memory Resource Controller disabled by default"
755 depends on CGROUP_MEM_RES_CTLR
758 Disable the memory group resource controller unless explicitly
759 enabled using the kernel parameter "cgroup_enable=memory".
761 config CGROUP_MEM_RES_CTLR_SWAP
762 bool "Memory Resource Controller Swap Extension"
763 depends on CGROUP_MEM_RES_CTLR && SWAP
765 Add swap management feature to memory resource controller. When you
766 enable this, you can limit mem+swap usage per cgroup. In other words,
767 when you disable this, memory resource controller has no cares to
768 usage of swap...a process can exhaust all of the swap. This extension
769 is useful when you want to avoid exhaustion swap but this itself
770 adds more overheads and consumes memory for remembering information.
771 Especially if you use 32bit system or small memory system, please
772 be careful about enabling this. When memory resource controller
773 is disabled by boot option, this will be automatically disabled and
774 there will be no overhead from this. Even when you set this config=y,
775 if boot option "swapaccount=0" is set, swap will not be accounted.
776 Now, memory usage of swap_cgroup is 2 bytes per entry. If swap page
777 size is 4096bytes, 512k per 1Gbytes of swap.
778 config CGROUP_MEM_RES_CTLR_SWAP_ENABLED
779 bool "Memory Resource Controller Swap Extension enabled by default"
780 depends on CGROUP_MEM_RES_CTLR_SWAP
783 Memory Resource Controller Swap Extension comes with its price in
784 a bigger memory consumption. General purpose distribution kernels
785 which want to enable the feature but keep it disabled by default
786 and let the user enable it by swapaccount boot command line
787 parameter should have this option unselected.
788 For those who want to have the feature enabled by default should
789 select this option (if, for some reason, they need to disable it
790 then swapaccount=0 does the trick).
791 config CGROUP_MEM_RES_CTLR_KMEM
792 bool "Memory Resource Controller Kernel Memory accounting (EXPERIMENTAL)"
793 depends on CGROUP_MEM_RES_CTLR && EXPERIMENTAL
796 The Kernel Memory extension for Memory Resource Controller can limit
797 the amount of memory used by kernel objects in the system. Those are
798 fundamentally different from the entities handled by the standard
799 Memory Controller, which are page-based, and can be swapped. Users of
800 the kmem extension can use it to guarantee that no group of processes
801 will ever exhaust kernel resources alone.
804 bool "Enable perf_event per-cpu per-container group (cgroup) monitoring"
805 depends on PERF_EVENTS && CGROUPS
807 This option extends the per-cpu mode to restrict monitoring to
808 threads which belong to the cgroup specified and run on the
813 menuconfig CGROUP_SCHED
814 bool "Group CPU scheduler"
815 depends on !SCHED_BFS
818 This feature lets CPU scheduler recognize task groups and control CPU
819 bandwidth allocation to such task groups. It uses cgroups to group
823 config FAIR_GROUP_SCHED
824 bool "Group scheduling for SCHED_OTHER"
825 depends on CGROUP_SCHED
829 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
830 depends on EXPERIMENTAL
831 depends on FAIR_GROUP_SCHED
834 This option allows users to define CPU bandwidth rates (limits) for
835 tasks running within the fair group scheduler. Groups with no limit
836 set are considered to be unconstrained and will run with no
838 See tip/Documentation/scheduler/sched-bwc.txt for more information.
840 config RT_GROUP_SCHED
841 bool "Group scheduling for SCHED_RR/FIFO"
842 depends on EXPERIMENTAL
843 depends on CGROUP_SCHED
846 This feature lets you explicitly allocate real CPU bandwidth
847 to task groups. If enabled, it will also make it impossible to
848 schedule realtime tasks for non-root users until you allocate
849 realtime bandwidth for them.
850 See Documentation/scheduler/sched-rt-group.txt for more information.
855 tristate "Block IO controller"
859 Generic block IO controller cgroup interface. This is the common
860 cgroup interface which should be used by various IO controlling
863 Currently, CFQ IO scheduler uses it to recognize task groups and
864 control disk bandwidth allocation (proportional time slice allocation)
865 to such task groups. It is also used by bio throttling logic in
866 block layer to implement upper limit in IO rates on a device.
868 This option only enables generic Block IO controller infrastructure.
869 One needs to also enable actual IO controlling logic/policy. For
870 enabling proportional weight division of disk bandwidth in CFQ, set
871 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
872 CONFIG_BLK_DEV_THROTTLING=y.
874 See Documentation/cgroups/blkio-controller.txt for more information.
876 config DEBUG_BLK_CGROUP
877 bool "Enable Block IO controller debugging"
878 depends on BLK_CGROUP
881 Enable some debugging help. Currently it exports additional stat
882 files in a cgroup which can be useful for debugging.
886 config CHECKPOINT_RESTORE
887 bool "Checkpoint/restore support" if EXPERT
890 Enables additional kernel features in a sake of checkpoint/restore.
891 In particular it adds auxiliary prctl codes to setup process text,
892 data and heap segment sizes, and a few additional /proc filesystem
895 If unsure, say N here.
897 menuconfig NAMESPACES
898 bool "Namespaces support" if EXPERT
901 Provides the way to make tasks work with different objects using
902 the same id. For example same IPC id may refer to different objects
903 or same user id or pid may refer to different tasks when used in
904 different namespaces.
912 In this namespace tasks see different info provided with the
917 depends on (SYSVIPC || POSIX_MQUEUE)
920 In this namespace tasks work with IPC ids which correspond to
921 different IPC objects in different namespaces.
924 bool "User namespace (EXPERIMENTAL)"
925 depends on EXPERIMENTAL
928 This allows containers, i.e. vservers, to use user namespaces
929 to provide different user info for different servers.
933 bool "PID Namespaces"
936 Support process id namespaces. This allows having multiple
937 processes with the same pid as long as they are in different
938 pid namespaces. This is a building block of containers.
941 bool "Network namespace"
945 Allow user space to create what appear to be multiple instances
946 of the network stack.
950 config SCHED_AUTOGROUP
951 bool "Automatic process group scheduling"
952 depends on !SCHED_BFS
956 select FAIR_GROUP_SCHED
958 This option optimizes the scheduler for common desktop workloads by
959 automatically creating and populating task groups. This separation
960 of workloads isolates aggressive CPU burners (like build jobs) from
961 desktop applications. Task group autogeneration is currently based
967 config SYSFS_DEPRECATED
968 bool "Enable deprecated sysfs features to support old userspace tools"
972 This option adds code that switches the layout of the "block" class
973 devices, to not show up in /sys/class/block/, but only in
976 This switch is only active when the sysfs.deprecated=1 boot option is
977 passed or the SYSFS_DEPRECATED_V2 option is set.
979 This option allows new kernels to run on old distributions and tools,
980 which might get confused by /sys/class/block/. Since 2007/2008 all
981 major distributions and tools handle this just fine.
983 Recent distributions and userspace tools after 2009/2010 depend on
984 the existence of /sys/class/block/, and will not work with this
987 Only if you are using a new kernel on an old distribution, you might
990 config SYSFS_DEPRECATED_V2
991 bool "Enable deprecated sysfs features by default"
994 depends on SYSFS_DEPRECATED
996 Enable deprecated sysfs by default.
998 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1001 Only if you are using a new kernel on an old distribution, you might
1002 need to say Y here. Even then, odds are you would not need it
1003 enabled, you can always pass the boot option if absolutely necessary.
1006 bool "Kernel->user space relay support (formerly relayfs)"
1008 This option enables support for relay interface support in
1009 certain file systems (such as debugfs).
1010 It is designed to provide an efficient mechanism for tools and
1011 facilities to relay large amounts of data from kernel space to
1016 config BLK_DEV_INITRD
1017 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1018 depends on BROKEN || !FRV
1020 The initial RAM filesystem is a ramfs which is loaded by the
1021 boot loader (loadlin or lilo) and that is mounted as root
1022 before the normal boot procedure. It is typically used to
1023 load modules needed to mount the "real" root file system,
1024 etc. See <file:Documentation/initrd.txt> for details.
1026 If RAM disk support (BLK_DEV_RAM) is also included, this
1027 also enables initial RAM disk (initrd) support and adds
1028 15 Kbytes (more on some other architectures) to the kernel size.
1034 source "usr/Kconfig"
1038 config CC_OPTIMIZE_FOR_SIZE
1039 bool "Optimize for size"
1041 Enabling this option will pass "-Os" instead of "-O2" to gcc
1042 resulting in a smaller kernel.
1053 bool "Configure standard kernel features (expert users)"
1054 # Unhide debug options, to make the on-by-default options visible
1057 This option allows certain base kernel options and settings
1058 to be disabled or tweaked. This is for specialized
1059 environments which can tolerate a "non-standard" kernel.
1060 Only use this if you really know what you are doing.
1063 bool "Enable 16-bit UID system calls" if EXPERT
1064 depends on ARM || BLACKFIN || CRIS || FRV || H8300 || X86_32 || M68K || (S390 && !64BIT) || SUPERH || SPARC32 || (SPARC64 && COMPAT) || UML || (X86_64 && IA32_EMULATION)
1067 This enables the legacy 16-bit UID syscall wrappers.
1069 config SYSCTL_SYSCALL
1070 bool "Sysctl syscall support" if EXPERT
1071 depends on PROC_SYSCTL
1075 sys_sysctl uses binary paths that have been found challenging
1076 to properly maintain and use. The interface in /proc/sys
1077 using paths with ascii names is now the primary path to this
1080 Almost nothing using the binary sysctl interface so if you are
1081 trying to save some space it is probably safe to disable this,
1082 making your kernel marginally smaller.
1084 If unsure say N here.
1087 bool "Load all symbols for debugging/ksymoops" if EXPERT
1090 Say Y here to let the kernel print out symbolic crash information and
1091 symbolic stack backtraces. This increases the size of the kernel
1092 somewhat, as all symbols have to be loaded into the kernel image.
1095 bool "Include all symbols in kallsyms"
1096 depends on DEBUG_KERNEL && KALLSYMS
1098 Normally kallsyms only contains the symbols of functions for nicer
1099 OOPS messages and backtraces (i.e., symbols from the text and inittext
1100 sections). This is sufficient for most cases. And only in very rare
1101 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1102 names of variables from the data sections, etc).
1104 This option makes sure that all symbols are loaded into the kernel
1105 image (i.e., symbols from all sections) in cost of increased kernel
1106 size (depending on the kernel configuration, it may be 300KiB or
1107 something like this).
1109 Say N unless you really need all symbols.
1112 bool "Support for hot-pluggable devices" if EXPERT
1115 This option is provided for the case where no hotplug or uevent
1116 capabilities is wanted by the kernel. You should only consider
1117 disabling this option for embedded systems that do not use modules, a
1118 dynamic /dev tree, or dynamic device discovery. Just say Y.
1122 bool "Enable support for printk" if EXPERT
1124 This option enables normal printk support. Removing it
1125 eliminates most of the message strings from the kernel image
1126 and makes the kernel more or less silent. As this makes it
1127 very difficult to diagnose system problems, saying N here is
1128 strongly discouraged.
1131 bool "BUG() support" if EXPERT
1134 Disabling this option eliminates support for BUG and WARN, reducing
1135 the size of your kernel image and potentially quietly ignoring
1136 numerous fatal conditions. You should only consider disabling this
1137 option for embedded systems with no facilities for reporting errors.
1142 bool "Enable ELF core dumps" if EXPERT
1144 Enable support for generating core dumps. Disabling saves about 4k.
1147 config PCSPKR_PLATFORM
1148 bool "Enable PC-Speaker support" if EXPERT
1149 depends on HAVE_PCSPKR_PLATFORM
1153 This option allows to disable the internal PC-Speaker
1154 support, saving some memory.
1156 config HAVE_PCSPKR_PLATFORM
1161 bool "Enable full-sized data structures for core" if EXPERT
1163 Disabling this option reduces the size of miscellaneous core
1164 kernel data structures. This saves memory on small machines,
1165 but may reduce performance.
1168 bool "Enable futex support" if EXPERT
1172 Disabling this option will cause the kernel to be built without
1173 support for "fast userspace mutexes". The resulting kernel may not
1174 run glibc-based applications correctly.
1177 bool "Enable eventpoll support" if EXPERT
1181 Disabling this option will cause the kernel to be built without
1182 support for epoll family of system calls.
1185 bool "Enable signalfd() system call" if EXPERT
1189 Enable the signalfd() system call that allows to receive signals
1190 on a file descriptor.
1195 bool "Enable timerfd() system call" if EXPERT
1199 Enable the timerfd() system call that allows to receive timer
1200 events on a file descriptor.
1205 bool "Enable eventfd() system call" if EXPERT
1209 Enable the eventfd() system call that allows to receive both
1210 kernel notification (ie. KAIO) or userspace notifications.
1215 bool "Use full shmem filesystem" if EXPERT
1219 The shmem is an internal filesystem used to manage shared memory.
1220 It is backed by swap and manages resource limits. It is also exported
1221 to userspace as tmpfs if TMPFS is enabled. Disabling this
1222 option replaces shmem and tmpfs with the much simpler ramfs code,
1223 which may be appropriate on small systems without swap.
1226 bool "Enable AIO support" if EXPERT
1229 This option enables POSIX asynchronous I/O which may by used
1230 by some high performance threaded applications. Disabling
1231 this option saves about 7k.
1234 bool "Embedded system"
1237 This option should be enabled if compiling the kernel for
1238 an embedded system so certain expert options are available
1241 config HAVE_PERF_EVENTS
1244 See tools/perf/design.txt for details.
1246 config PERF_USE_VMALLOC
1249 See tools/perf/design.txt for details
1251 menu "Kernel Performance Events And Counters"
1254 bool "Kernel performance events and counters"
1255 default y if (PROFILING || PERF_COUNTERS)
1256 depends on HAVE_PERF_EVENTS
1260 Enable kernel support for various performance events provided
1261 by software and hardware.
1263 Software events are supported either built-in or via the
1264 use of generic tracepoints.
1266 Most modern CPUs support performance events via performance
1267 counter registers. These registers count the number of certain
1268 types of hw events: such as instructions executed, cachemisses
1269 suffered, or branches mis-predicted - without slowing down the
1270 kernel or applications. These registers can also trigger interrupts
1271 when a threshold number of events have passed - and can thus be
1272 used to profile the code that runs on that CPU.
1274 The Linux Performance Event subsystem provides an abstraction of
1275 these software and hardware event capabilities, available via a
1276 system call and used by the "perf" utility in tools/perf/. It
1277 provides per task and per CPU counters, and it provides event
1278 capabilities on top of those.
1282 config PERF_COUNTERS
1283 bool "Kernel performance counters (old config option)"
1284 depends on HAVE_PERF_EVENTS
1286 This config has been obsoleted by the PERF_EVENTS
1287 config option - please see that one for details.
1289 It has no effect on the kernel whether you enable
1290 it or not, it is a compatibility placeholder.
1294 config DEBUG_PERF_USE_VMALLOC
1296 bool "Debug: use vmalloc to back perf mmap() buffers"
1297 depends on PERF_EVENTS && DEBUG_KERNEL
1298 select PERF_USE_VMALLOC
1300 Use vmalloc memory to back perf mmap() buffers.
1302 Mostly useful for debugging the vmalloc code on platforms
1303 that don't require it.
1309 config VM_EVENT_COUNTERS
1311 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1313 VM event counters are needed for event counts to be shown.
1314 This option allows the disabling of the VM event counters
1315 on EXPERT systems. /proc/vmstat will only show page counts
1316 if VM event counters are disabled.
1320 bool "Enable PCI quirk workarounds" if EXPERT
1323 This enables workarounds for various PCI chipset
1324 bugs/quirks. Disable this only if your target machine is
1325 unaffected by PCI quirks.
1329 bool "Enable SLUB debugging support" if EXPERT
1330 depends on SLUB && SYSFS
1332 SLUB has extensive debug support features. Disabling these can
1333 result in significant savings in code size. This also disables
1334 SLUB sysfs support. /sys/slab will not exist and there will be
1335 no support for cache validation etc.
1338 bool "Disable heap randomization"
1341 Randomizing heap placement makes heap exploits harder, but it
1342 also breaks ancient binaries (including anything libc5 based).
1343 This option changes the bootup default to heap randomization
1344 disabled, and can be overridden at runtime by setting
1345 /proc/sys/kernel/randomize_va_space to 2.
1347 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1350 prompt "Choose SLAB allocator"
1353 This option allows to select a slab allocator.
1358 The regular slab allocator that is established and known to work
1359 well in all environments. It organizes cache hot objects in
1360 per cpu and per node queues.
1363 bool "SLUB (Unqueued Allocator)"
1365 SLUB is a slab allocator that minimizes cache line usage
1366 instead of managing queues of cached objects (SLAB approach).
1367 Per cpu caching is realized using slabs of objects instead
1368 of queues of objects. SLUB can use memory efficiently
1369 and has enhanced diagnostics. SLUB is the default choice for
1374 bool "SLOB (Simple Allocator)"
1376 SLOB replaces the stock allocator with a drastically simpler
1377 allocator. SLOB is generally more space efficient but
1378 does not perform as well on large systems.
1382 config MMAP_ALLOW_UNINITIALIZED
1383 bool "Allow mmapped anonymous memory to be uninitialized"
1384 depends on EXPERT && !MMU
1387 Normally, and according to the Linux spec, anonymous memory obtained
1388 from mmap() has it's contents cleared before it is passed to
1389 userspace. Enabling this config option allows you to request that
1390 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1391 providing a huge performance boost. If this option is not enabled,
1392 then the flag will be ignored.
1394 This is taken advantage of by uClibc's malloc(), and also by
1395 ELF-FDPIC binfmt's brk and stack allocator.
1397 Because of the obvious security issues, this option should only be
1398 enabled on embedded devices where you control what is run in
1399 userspace. Since that isn't generally a problem on no-MMU systems,
1400 it is normally safe to say Y here.
1402 See Documentation/nommu-mmap.txt for more information.
1405 bool "Profiling support"
1407 Say Y here to enable the extended profiling support mechanisms used
1408 by profilers such as OProfile.
1411 # Place an empty function call at each tracepoint site. Can be
1412 # dynamically changed for a probe function.
1417 source "arch/Kconfig"
1419 endmenu # General setup
1421 config HAVE_GENERIC_DMA_COHERENT
1428 depends on SLAB || SLUB_DEBUG
1436 default 0 if BASE_FULL
1437 default 1 if !BASE_FULL
1440 bool "Enable loadable module support"
1442 Kernel modules are small pieces of compiled code which can
1443 be inserted in the running kernel, rather than being
1444 permanently built into the kernel. You use the "modprobe"
1445 tool to add (and sometimes remove) them. If you say Y here,
1446 many parts of the kernel can be built as modules (by
1447 answering M instead of Y where indicated): this is most
1448 useful for infrequently used options which are not required
1449 for booting. For more information, see the man pages for
1450 modprobe, lsmod, modinfo, insmod and rmmod.
1452 If you say Y here, you will need to run "make
1453 modules_install" to put the modules under /lib/modules/
1454 where modprobe can find them (you may need to be root to do
1461 config MODULE_FORCE_LOAD
1462 bool "Forced module loading"
1465 Allow loading of modules without version information (ie. modprobe
1466 --force). Forced module loading sets the 'F' (forced) taint flag and
1467 is usually a really bad idea.
1469 config MODULE_UNLOAD
1470 bool "Module unloading"
1472 Without this option you will not be able to unload any
1473 modules (note that some modules may not be unloadable
1474 anyway), which makes your kernel smaller, faster
1475 and simpler. If unsure, say Y.
1477 config MODULE_FORCE_UNLOAD
1478 bool "Forced module unloading"
1479 depends on MODULE_UNLOAD && EXPERIMENTAL
1481 This option allows you to force a module to unload, even if the
1482 kernel believes it is unsafe: the kernel will remove the module
1483 without waiting for anyone to stop using it (using the -f option to
1484 rmmod). This is mainly for kernel developers and desperate users.
1488 bool "Module versioning support"
1490 Usually, you have to use modules compiled with your kernel.
1491 Saying Y here makes it sometimes possible to use modules
1492 compiled for different kernels, by adding enough information
1493 to the modules to (hopefully) spot any changes which would
1494 make them incompatible with the kernel you are running. If
1497 config MODULE_SRCVERSION_ALL
1498 bool "Source checksum for all modules"
1500 Modules which contain a MODULE_VERSION get an extra "srcversion"
1501 field inserted into their modinfo section, which contains a
1502 sum of the source files which made it. This helps maintainers
1503 see exactly which source was used to build a module (since
1504 others sometimes change the module source without updating
1505 the version). With this option, such a "srcversion" field
1506 will be created for all modules. If unsure, say N.
1510 config INIT_ALL_POSSIBLE
1513 Back when each arch used to define their own cpu_online_map and
1514 cpu_possible_map, some of them chose to initialize cpu_possible_map
1515 with all 1s, and others with all 0s. When they were centralised,
1516 it was better to provide this option than to break all the archs
1517 and have several arch maintainers pursuing me down dark alleys.
1522 depends on (SMP && MODULE_UNLOAD) || HOTPLUG_CPU
1524 Need stop_machine() primitive.
1526 source "block/Kconfig"
1528 config PREEMPT_NOTIFIERS
1535 source "kernel/Kconfig.locks"