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"
27 bool "Prompt for development and/or incomplete code/drivers"
29 Some of the various things that Linux supports (such as network
30 drivers, file systems, network protocols, etc.) can be in a state
31 of development where the functionality, stability, or the level of
32 testing is not yet high enough for general use. This is usually
33 known as the "alpha-test" phase among developers. If a feature is
34 currently in alpha-test, then the developers usually discourage
35 uninformed widespread use of this feature by the general public to
36 avoid "Why doesn't this work?" type mail messages. However, active
37 testing and use of these systems is welcomed. Just be aware that it
38 may not meet the normal level of reliability or it may fail to work
39 in some special cases. Detailed bug reports from people familiar
40 with the kernel internals are usually welcomed by the developers
41 (before submitting bug reports, please read the documents
42 <file:README>, <file:MAINTAINERS>, <file:REPORTING-BUGS>,
43 <file:Documentation/BUG-HUNTING>, and
44 <file:Documentation/oops-tracing.txt> in the kernel source).
46 This option will also make obsoleted drivers available. These are
47 drivers that have been replaced by something else, and/or are
48 scheduled to be removed in a future kernel release.
50 Unless you intend to help test and develop a feature or driver that
51 falls into this category, or you have a situation that requires
52 using these features, you should probably say N here, which will
53 cause the configurator to present you with fewer choices. If
54 you say Y here, you will be offered the choice of using features or
55 drivers that are currently considered to be in the alpha-test phase.
62 depends on BROKEN || !SMP
67 depends on SMP || PREEMPT
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 "Local version - append to kernel release"
82 Append an extra string to the end of your kernel version.
83 This will show up when you type uname, for example.
84 The string you set here will be appended after the contents of
85 any files with a filename matching localversion* in your
86 object and source tree, in that order. Your total string can
87 be a maximum of 64 characters.
89 config LOCALVERSION_AUTO
90 bool "Automatically append version information to the version string"
93 This will try to automatically determine if the current tree is a
94 release tree by looking for git tags that belong to the current
97 A string of the format -gxxxxxxxx will be added to the localversion
98 if a git-based tree is found. The string generated by this will be
99 appended after any matching localversion* files, and after the value
100 set in CONFIG_LOCALVERSION.
102 (The actual string used here is the first eight characters produced
103 by running the command:
105 $ git rev-parse --verify HEAD
107 which is done within the script "scripts/setlocalversion".)
109 config HAVE_KERNEL_GZIP
112 config HAVE_KERNEL_BZIP2
115 config HAVE_KERNEL_LZMA
118 config HAVE_KERNEL_LZO
122 prompt "Kernel compression mode"
124 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_LZO
126 The linux kernel is a kind of self-extracting executable.
127 Several compression algorithms are available, which differ
128 in efficiency, compression and decompression speed.
129 Compression speed is only relevant when building a kernel.
130 Decompression speed is relevant at each boot.
132 If you have any problems with bzip2 or lzma compressed
133 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
134 version of this functionality (bzip2 only), for 2.4, was
135 supplied by Christian Ludwig)
137 High compression options are mostly useful for users, who
138 are low on disk space (embedded systems), but for whom ram
141 If in doubt, select 'gzip'
145 depends on HAVE_KERNEL_GZIP
147 The old and tried gzip compression. It provides a good balance
148 between compression ratio and decompression speed.
152 depends on HAVE_KERNEL_BZIP2
154 Its compression ratio and speed is intermediate.
155 Decompression speed is slowest among the three. The kernel
156 size is about 10% smaller with bzip2, in comparison to gzip.
157 Bzip2 uses a large amount of memory. For modern kernels you
158 will need at least 8MB RAM or more for booting.
162 depends on HAVE_KERNEL_LZMA
164 The most recent compression algorithm.
165 Its ratio is best, decompression speed is between the other
166 two. Compression is slowest. The kernel size is about 33%
167 smaller with LZMA in comparison to gzip.
171 depends on HAVE_KERNEL_LZO
173 Its compression ratio is the poorest among the 4. The kernel
174 size is about about 10% bigger than gzip; however its speed
175 (both compression and decompression) is the fastest.
180 bool "Support for paging of anonymous memory (swap)"
181 depends on MMU && BLOCK
184 This option allows you to choose whether you want to have support
185 for so called swap devices or swap files in your kernel that are
186 used to provide more virtual memory than the actual RAM present
187 in your computer. If unsure say Y.
192 Inter Process Communication is a suite of library functions and
193 system calls which let processes (running programs) synchronize and
194 exchange information. It is generally considered to be a good thing,
195 and some programs won't run unless you say Y here. In particular, if
196 you want to run the DOS emulator dosemu under Linux (read the
197 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
198 you'll need to say Y here.
200 You can find documentation about IPC with "info ipc" and also in
201 section 6.4 of the Linux Programmer's Guide, available from
202 <http://www.tldp.org/guides.html>.
204 config SYSVIPC_SYSCTL
211 bool "POSIX Message Queues"
212 depends on NET && EXPERIMENTAL
214 POSIX variant of message queues is a part of IPC. In POSIX message
215 queues every message has a priority which decides about succession
216 of receiving it by a process. If you want to compile and run
217 programs written e.g. for Solaris with use of its POSIX message
218 queues (functions mq_*) say Y here.
220 POSIX message queues are visible as a filesystem called 'mqueue'
221 and can be mounted somewhere if you want to do filesystem
222 operations on message queues.
226 config POSIX_MQUEUE_SYSCTL
228 depends on POSIX_MQUEUE
232 config BSD_PROCESS_ACCT
233 bool "BSD Process Accounting"
235 If you say Y here, a user level program will be able to instruct the
236 kernel (via a special system call) to write process accounting
237 information to a file: whenever a process exits, information about
238 that process will be appended to the file by the kernel. The
239 information includes things such as creation time, owning user,
240 command name, memory usage, controlling terminal etc. (the complete
241 list is in the struct acct in <file:include/linux/acct.h>). It is
242 up to the user level program to do useful things with this
243 information. This is generally a good idea, so say Y.
245 config BSD_PROCESS_ACCT_V3
246 bool "BSD Process Accounting version 3 file format"
247 depends on BSD_PROCESS_ACCT
250 If you say Y here, the process accounting information is written
251 in a new file format that also logs the process IDs of each
252 process and it's parent. Note that this file format is incompatible
253 with previous v0/v1/v2 file formats, so you will need updated tools
254 for processing it. A preliminary version of these tools is available
255 at <http://www.gnu.org/software/acct/>.
258 bool "Export task/process statistics through netlink (EXPERIMENTAL)"
262 Export selected statistics for tasks/processes through the
263 generic netlink interface. Unlike BSD process accounting, the
264 statistics are available during the lifetime of tasks/processes as
265 responses to commands. Like BSD accounting, they are sent to user
270 config TASK_DELAY_ACCT
271 bool "Enable per-task delay accounting (EXPERIMENTAL)"
274 Collect information on time spent by a task waiting for system
275 resources like cpu, synchronous block I/O completion and swapping
276 in pages. Such statistics can help in setting a task's priorities
277 relative to other tasks for cpu, io, rss limits etc.
282 bool "Enable extended accounting over taskstats (EXPERIMENTAL)"
285 Collect extended task accounting data and send the data
286 to userland for processing over the taskstats interface.
290 config TASK_IO_ACCOUNTING
291 bool "Enable per-task storage I/O accounting (EXPERIMENTAL)"
292 depends on TASK_XACCT
294 Collect information on the number of bytes of storage I/O which this
300 bool "Auditing support"
303 Enable auditing infrastructure that can be used with another
304 kernel subsystem, such as SELinux (which requires this for
305 logging of avc messages output). Does not do system-call
306 auditing without CONFIG_AUDITSYSCALL.
309 bool "Enable system-call auditing support"
310 depends on AUDIT && (X86 || PPC || S390 || IA64 || UML || SPARC64 || SUPERH)
311 default y if SECURITY_SELINUX
313 Enable low-overhead system-call auditing infrastructure that
314 can be used independently or with another kernel subsystem,
315 such as SELinux. To use audit's filesystem watch feature, please
316 ensure that INOTIFY is configured.
320 depends on AUDITSYSCALL
326 prompt "RCU Implementation"
330 bool "Tree-based hierarchical RCU"
332 This option selects the RCU implementation that is
333 designed for very large SMP system with hundreds or
334 thousands of CPUs. It also scales down nicely to
337 config TREE_PREEMPT_RCU
338 bool "Preemptable tree-based hierarchical RCU"
341 This option selects the RCU implementation that is
342 designed for very large SMP systems with hundreds or
343 thousands of CPUs, but for which real-time response
344 is also required. It also scales down nicely to
348 bool "UP-only small-memory-footprint RCU"
351 This option selects the RCU implementation that is
352 designed for UP systems from which real-time response
353 is not required. This option greatly reduces the
354 memory footprint of RCU.
359 bool "Enable tracing for RCU"
360 depends on TREE_RCU || TREE_PREEMPT_RCU
362 This option provides tracing in RCU which presents stats
363 in debugfs for debugging RCU implementation.
365 Say Y here if you want to enable RCU tracing
366 Say N if you are unsure.
369 int "Tree-based hierarchical RCU fanout value"
372 depends on TREE_RCU || TREE_PREEMPT_RCU
376 This option controls the fanout of hierarchical implementations
377 of RCU, allowing RCU to work efficiently on machines with
378 large numbers of CPUs. This value must be at least the cube
379 root of NR_CPUS, which allows NR_CPUS up to 32,768 for 32-bit
380 systems and up to 262,144 for 64-bit systems.
382 Select a specific number if testing RCU itself.
383 Take the default if unsure.
385 config RCU_FANOUT_EXACT
386 bool "Disable tree-based hierarchical RCU auto-balancing"
387 depends on TREE_RCU || TREE_PREEMPT_RCU
390 This option forces use of the exact RCU_FANOUT value specified,
391 regardless of imbalances in the hierarchy. This is useful for
392 testing RCU itself, and might one day be useful on systems with
393 strong NUMA behavior.
395 Without RCU_FANOUT_EXACT, the code will balance the hierarchy.
399 config TREE_RCU_TRACE
400 def_bool RCU_TRACE && ( TREE_RCU || TREE_PREEMPT_RCU )
403 This option provides tracing for the TREE_RCU and
404 TREE_PREEMPT_RCU implementations, permitting Makefile to
405 trivially select kernel/rcutree_trace.c.
407 endmenu # "RCU Subsystem"
410 tristate "Kernel .config support"
412 This option enables the complete Linux kernel ".config" file
413 contents to be saved in the kernel. It provides documentation
414 of which kernel options are used in a running kernel or in an
415 on-disk kernel. This information can be extracted from the kernel
416 image file with the script scripts/extract-ikconfig and used as
417 input to rebuild the current kernel or to build another kernel.
418 It can also be extracted from a running kernel by reading
419 /proc/config.gz if enabled (below).
422 bool "Enable access to .config through /proc/config.gz"
423 depends on IKCONFIG && PROC_FS
425 This option enables access to the kernel configuration file
426 through /proc/config.gz.
429 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
433 Select kernel log buffer size as a power of 2.
443 # Architectures with an unreliable sched_clock() should select this:
445 config HAVE_UNSTABLE_SCHED_CLOCK
449 bool "Group CPU scheduler"
450 depends on EXPERIMENTAL
453 This feature lets CPU scheduler recognize task groups and control CPU
454 bandwidth allocation to such task groups.
455 In order to create a group from arbitrary set of processes, use
456 CONFIG_CGROUPS. (See Control Group support.)
458 config FAIR_GROUP_SCHED
459 bool "Group scheduling for SCHED_OTHER"
460 depends on GROUP_SCHED
463 config RT_GROUP_SCHED
464 bool "Group scheduling for SCHED_RR/FIFO"
465 depends on EXPERIMENTAL
466 depends on GROUP_SCHED
469 This feature lets you explicitly allocate real CPU bandwidth
470 to users or control groups (depending on the "Basis for grouping tasks"
471 setting below. If enabled, it will also make it impossible to
472 schedule realtime tasks for non-root users until you allocate
473 realtime bandwidth for them.
474 See Documentation/scheduler/sched-rt-group.txt for more information.
477 depends on GROUP_SCHED
478 prompt "Basis for grouping tasks"
484 This option will choose userid as the basis for grouping
485 tasks, thus providing equal CPU bandwidth to each user.
488 bool "Control groups"
491 This option allows you to create arbitrary task groups
492 using the "cgroup" pseudo filesystem and control
493 the cpu bandwidth allocated to each such task group.
494 Refer to Documentation/cgroups/cgroups.txt for more
495 information on "cgroup" pseudo filesystem.
500 boolean "Control Group support"
502 This option adds support for grouping sets of processes together, for
503 use with process control subsystems such as Cpusets, CFS, memory
504 controls or device isolation.
506 - Documentation/scheduler/sched-design-CFS.txt (CFS)
507 - Documentation/cgroups/ (features for grouping, isolation
508 and resource control)
515 bool "Example debug cgroup subsystem"
519 This option enables a simple cgroup subsystem that
520 exports useful debugging information about the cgroups
526 bool "Namespace cgroup subsystem"
529 Provides a simple namespace cgroup subsystem to
530 provide hierarchical naming of sets of namespaces,
531 for instance virtual servers and checkpoint/restart
534 config CGROUP_FREEZER
535 bool "Freezer cgroup subsystem"
538 Provides a way to freeze and unfreeze all tasks in a
542 bool "Device controller for cgroups"
543 depends on CGROUPS && EXPERIMENTAL
545 Provides a cgroup implementing whitelists for devices which
546 a process in the cgroup can mknod or open.
549 bool "Cpuset support"
552 This option will let you create and manage CPUSETs which
553 allow dynamically partitioning a system into sets of CPUs and
554 Memory Nodes and assigning tasks to run only within those sets.
555 This is primarily useful on large SMP or NUMA systems.
559 config PROC_PID_CPUSET
560 bool "Include legacy /proc/<pid>/cpuset file"
564 config CGROUP_CPUACCT
565 bool "Simple CPU accounting cgroup subsystem"
568 Provides a simple Resource Controller for monitoring the
569 total CPU consumed by the tasks in a cgroup.
571 config RESOURCE_COUNTERS
572 bool "Resource counters"
574 This option enables controller independent resource accounting
575 infrastructure that works with cgroups.
578 config CGROUP_MEM_RES_CTLR
579 bool "Memory Resource Controller for Control Groups"
580 depends on CGROUPS && RESOURCE_COUNTERS
583 Provides a memory resource controller that manages both anonymous
584 memory and page cache. (See Documentation/cgroups/memory.txt)
586 Note that setting this option increases fixed memory overhead
587 associated with each page of memory in the system. By this,
588 20(40)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory
589 usage tracking struct at boot. Total amount of this is printed out
592 Only enable when you're ok with these trade offs and really
593 sure you need the memory resource controller. Even when you enable
594 this, you can set "cgroup_disable=memory" at your boot option to
595 disable memory resource controller and you can avoid overheads.
596 (and lose benefits of memory resource controller)
598 This config option also selects MM_OWNER config option, which
599 could in turn add some fork/exit overhead.
601 config CGROUP_MEM_RES_CTLR_SWAP
602 bool "Memory Resource Controller Swap Extension(EXPERIMENTAL)"
603 depends on CGROUP_MEM_RES_CTLR && SWAP && EXPERIMENTAL
605 Add swap management feature to memory resource controller. When you
606 enable this, you can limit mem+swap usage per cgroup. In other words,
607 when you disable this, memory resource controller has no cares to
608 usage of swap...a process can exhaust all of the swap. This extension
609 is useful when you want to avoid exhaustion swap but this itself
610 adds more overheads and consumes memory for remembering information.
611 Especially if you use 32bit system or small memory system, please
612 be careful about enabling this. When memory resource controller
613 is disabled by boot option, this will be automatically disabled and
614 there will be no overhead from this. Even when you set this config=y,
615 if boot option "noswapaccount" is set, swap will not be accounted.
616 Now, memory usage of swap_cgroup is 2 bytes per entry. If swap page
617 size is 4096bytes, 512k per 1Gbytes of swap.
624 config SYSFS_DEPRECATED
627 config SYSFS_DEPRECATED_V2
628 bool "enable deprecated sysfs features to support old userspace tools"
631 select SYSFS_DEPRECATED
633 This option switches the layout of sysfs to the deprecated
634 version. Do not use it on recent distributions.
636 The current sysfs layout features a unified device tree at
637 /sys/devices/, which is able to express a hierarchy between
638 class devices. If the deprecated option is set to Y, the
639 unified device tree is split into a bus device tree at
640 /sys/devices/ and several individual class device trees at
641 /sys/class/. The class and bus devices will be connected by
642 "<subsystem>:<name>" and the "device" links. The "block"
643 class devices, will not show up in /sys/class/block/. Some
644 subsystems will suppress the creation of some devices which
645 depend on the unified device tree.
647 This option is not a pure compatibility option that can
648 be safely enabled on newer distributions. It will change the
649 layout of sysfs to the non-extensible deprecated version,
650 and disable some features, which can not be exported without
651 confusing older userspace tools. Since 2007/2008 all major
652 distributions do not enable this option, and ship no tools which
653 depend on the deprecated layout or this option.
655 If you are using a new kernel on an older distribution, or use
656 older userspace tools, you might need to say Y here. Do not say Y,
657 if the original kernel, that came with your distribution, has
658 this option set to N.
661 bool "Kernel->user space relay support (formerly relayfs)"
663 This option enables support for relay interface support in
664 certain file systems (such as debugfs).
665 It is designed to provide an efficient mechanism for tools and
666 facilities to relay large amounts of data from kernel space to
672 bool "Namespaces support" if EMBEDDED
675 Provides the way to make tasks work with different objects using
676 the same id. For example same IPC id may refer to different objects
677 or same user id or pid may refer to different tasks when used in
678 different namespaces.
682 depends on NAMESPACES
684 In this namespace tasks see different info provided with the
689 depends on NAMESPACES && (SYSVIPC || POSIX_MQUEUE)
691 In this namespace tasks work with IPC ids which correspond to
692 different IPC objects in different namespaces.
695 bool "User namespace (EXPERIMENTAL)"
696 depends on NAMESPACES && EXPERIMENTAL
698 This allows containers, i.e. vservers, to use user namespaces
699 to provide different user info for different servers.
703 bool "PID Namespaces (EXPERIMENTAL)"
705 depends on NAMESPACES && EXPERIMENTAL
707 Support process id namespaces. This allows having multiple
708 processes with the same pid as long as they are in different
709 pid namespaces. This is a building block of containers.
711 Unless you want to work with an experimental feature
715 bool "Network namespace"
717 depends on NAMESPACES && EXPERIMENTAL && NET
719 Allow user space to create what appear to be multiple instances
720 of the network stack.
722 config BLK_DEV_INITRD
723 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
724 depends on BROKEN || !FRV
726 The initial RAM filesystem is a ramfs which is loaded by the
727 boot loader (loadlin or lilo) and that is mounted as root
728 before the normal boot procedure. It is typically used to
729 load modules needed to mount the "real" root file system,
730 etc. See <file:Documentation/initrd.txt> for details.
732 If RAM disk support (BLK_DEV_RAM) is also included, this
733 also enables initial RAM disk (initrd) support and adds
734 15 Kbytes (more on some other architectures) to the kernel size.
744 config CC_OPTIMIZE_FOR_SIZE
745 bool "Optimize for size"
748 Enabling this option will pass "-Os" instead of "-O2" to gcc
749 resulting in a smaller kernel.
760 bool "Configure standard kernel features (for small systems)"
762 This option allows certain base kernel options and settings
763 to be disabled or tweaked. This is for specialized
764 environments which can tolerate a "non-standard" kernel.
765 Only use this if you really know what you are doing.
768 bool "Enable 16-bit UID system calls" if EMBEDDED
769 depends on ARM || BLACKFIN || CRIS || FRV || H8300 || X86_32 || M68K || (S390 && !64BIT) || SUPERH || SPARC32 || (SPARC64 && COMPAT) || UML || (X86_64 && IA32_EMULATION)
772 This enables the legacy 16-bit UID syscall wrappers.
774 config SYSCTL_SYSCALL
775 bool "Sysctl syscall support" if EMBEDDED
776 depends on PROC_SYSCTL
780 sys_sysctl uses binary paths that have been found challenging
781 to properly maintain and use. The interface in /proc/sys
782 using paths with ascii names is now the primary path to this
785 Almost nothing using the binary sysctl interface so if you are
786 trying to save some space it is probably safe to disable this,
787 making your kernel marginally smaller.
789 If unsure say Y here.
792 bool "Load all symbols for debugging/ksymoops" if EMBEDDED
795 Say Y here to let the kernel print out symbolic crash information and
796 symbolic stack backtraces. This increases the size of the kernel
797 somewhat, as all symbols have to be loaded into the kernel image.
800 bool "Include all symbols in kallsyms"
801 depends on DEBUG_KERNEL && KALLSYMS
803 Normally kallsyms only contains the symbols of functions, for nicer
804 OOPS messages. Some debuggers can use kallsyms for other
805 symbols too: say Y here to include all symbols, if you need them
806 and you don't care about adding 300k to the size of your kernel.
810 config KALLSYMS_EXTRA_PASS
811 bool "Do an extra kallsyms pass"
814 If kallsyms is not working correctly, the build will fail with
815 inconsistent kallsyms data. If that occurs, log a bug report and
816 turn on KALLSYMS_EXTRA_PASS which should result in a stable build.
817 Always say N here unless you find a bug in kallsyms, which must be
818 reported. KALLSYMS_EXTRA_PASS is only a temporary workaround while
819 you wait for kallsyms to be fixed.
823 bool "Support for hot-pluggable devices" if EMBEDDED
826 This option is provided for the case where no hotplug or uevent
827 capabilities is wanted by the kernel. You should only consider
828 disabling this option for embedded systems that do not use modules, a
829 dynamic /dev tree, or dynamic device discovery. Just say Y.
833 bool "Enable support for printk" if EMBEDDED
835 This option enables normal printk support. Removing it
836 eliminates most of the message strings from the kernel image
837 and makes the kernel more or less silent. As this makes it
838 very difficult to diagnose system problems, saying N here is
839 strongly discouraged.
842 bool "BUG() support" if EMBEDDED
845 Disabling this option eliminates support for BUG and WARN, reducing
846 the size of your kernel image and potentially quietly ignoring
847 numerous fatal conditions. You should only consider disabling this
848 option for embedded systems with no facilities for reporting errors.
853 bool "Enable ELF core dumps" if EMBEDDED
855 Enable support for generating core dumps. Disabling saves about 4k.
857 config PCSPKR_PLATFORM
858 bool "Enable PC-Speaker support" if EMBEDDED
859 depends on ALPHA || X86 || MIPS || PPC_PREP || PPC_CHRP || PPC_PSERIES
862 This option allows to disable the internal PC-Speaker
863 support, saving some memory.
867 bool "Enable full-sized data structures for core" if EMBEDDED
869 Disabling this option reduces the size of miscellaneous core
870 kernel data structures. This saves memory on small machines,
871 but may reduce performance.
874 bool "Enable futex support" if EMBEDDED
878 Disabling this option will cause the kernel to be built without
879 support for "fast userspace mutexes". The resulting kernel may not
880 run glibc-based applications correctly.
883 bool "Enable eventpoll support" if EMBEDDED
887 Disabling this option will cause the kernel to be built without
888 support for epoll family of system calls.
891 bool "Enable signalfd() system call" if EMBEDDED
895 Enable the signalfd() system call that allows to receive signals
896 on a file descriptor.
901 bool "Enable timerfd() system call" if EMBEDDED
905 Enable the timerfd() system call that allows to receive timer
906 events on a file descriptor.
911 bool "Enable eventfd() system call" if EMBEDDED
915 Enable the eventfd() system call that allows to receive both
916 kernel notification (ie. KAIO) or userspace notifications.
921 bool "Use full shmem filesystem" if EMBEDDED
925 The shmem is an internal filesystem used to manage shared memory.
926 It is backed by swap and manages resource limits. It is also exported
927 to userspace as tmpfs if TMPFS is enabled. Disabling this
928 option replaces shmem and tmpfs with the much simpler ramfs code,
929 which may be appropriate on small systems without swap.
932 bool "Enable AIO support" if EMBEDDED
935 This option enables POSIX asynchronous I/O which may by used
936 by some high performance threaded applications. Disabling
937 this option saves about 7k.
939 config HAVE_PERF_EVENTS
942 See tools/perf/design.txt for details.
944 config PERF_USE_VMALLOC
947 See tools/perf/design.txt for details
949 menu "Kernel Performance Events And Counters"
952 bool "Kernel performance events and counters"
953 default y if (PROFILING || PERF_COUNTERS)
954 depends on HAVE_PERF_EVENTS
957 Enable kernel support for various performance events provided
958 by software and hardware.
960 Software events are supported either built-in or via the
961 use of generic tracepoints.
963 Most modern CPUs support performance events via performance
964 counter registers. These registers count the number of certain
965 types of hw events: such as instructions executed, cachemisses
966 suffered, or branches mis-predicted - without slowing down the
967 kernel or applications. These registers can also trigger interrupts
968 when a threshold number of events have passed - and can thus be
969 used to profile the code that runs on that CPU.
971 The Linux Performance Event subsystem provides an abstraction of
972 these software and hardware event capabilities, available via a
973 system call and used by the "perf" utility in tools/perf/. It
974 provides per task and per CPU counters, and it provides event
975 capabilities on top of those.
980 bool "Tracepoint profiling sources"
981 depends on PERF_EVENTS && EVENT_TRACING
984 Allow the use of tracepoints as software performance events.
986 When this is enabled, you can create perf events based on
987 tracepoints using PERF_TYPE_TRACEPOINT and the tracepoint ID
988 found in debugfs://tracing/events/*/*/id. (The -e/--events
989 option to the perf tool can parse and interpret symbolic
990 tracepoints, in the subsystem:tracepoint_name format.)
993 bool "Kernel performance counters (old config option)"
994 depends on HAVE_PERF_EVENTS
996 This config has been obsoleted by the PERF_EVENTS
997 config option - please see that one for details.
999 It has no effect on the kernel whether you enable
1000 it or not, it is a compatibility placeholder.
1004 config DEBUG_PERF_USE_VMALLOC
1006 bool "Debug: use vmalloc to back perf mmap() buffers"
1007 depends on PERF_EVENTS && DEBUG_KERNEL
1008 select PERF_USE_VMALLOC
1010 Use vmalloc memory to back perf mmap() buffers.
1012 Mostly useful for debugging the vmalloc code on platforms
1013 that don't require it.
1019 config VM_EVENT_COUNTERS
1021 bool "Enable VM event counters for /proc/vmstat" if EMBEDDED
1023 VM event counters are needed for event counts to be shown.
1024 This option allows the disabling of the VM event counters
1025 on EMBEDDED systems. /proc/vmstat will only show page counts
1026 if VM event counters are disabled.
1030 bool "Enable PCI quirk workarounds" if EMBEDDED
1033 This enables workarounds for various PCI chipset
1034 bugs/quirks. Disable this only if your target machine is
1035 unaffected by PCI quirks.
1039 bool "Enable SLUB debugging support" if EMBEDDED
1040 depends on SLUB && SYSFS
1042 SLUB has extensive debug support features. Disabling these can
1043 result in significant savings in code size. This also disables
1044 SLUB sysfs support. /sys/slab will not exist and there will be
1045 no support for cache validation etc.
1048 bool "Disable heap randomization"
1051 Randomizing heap placement makes heap exploits harder, but it
1052 also breaks ancient binaries (including anything libc5 based).
1053 This option changes the bootup default to heap randomization
1054 disabled, and can be overridden at runtime by setting
1055 /proc/sys/kernel/randomize_va_space to 2.
1057 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1060 prompt "Choose SLAB allocator"
1063 This option allows to select a slab allocator.
1068 The regular slab allocator that is established and known to work
1069 well in all environments. It organizes cache hot objects in
1070 per cpu and per node queues.
1073 bool "SLUB (Unqueued Allocator)"
1075 SLUB is a slab allocator that minimizes cache line usage
1076 instead of managing queues of cached objects (SLAB approach).
1077 Per cpu caching is realized using slabs of objects instead
1078 of queues of objects. SLUB can use memory efficiently
1079 and has enhanced diagnostics. SLUB is the default choice for
1084 bool "SLOB (Simple Allocator)"
1086 SLOB replaces the stock allocator with a drastically simpler
1087 allocator. SLOB is generally more space efficient but
1088 does not perform as well on large systems.
1092 config MMAP_ALLOW_UNINITIALIZED
1093 bool "Allow mmapped anonymous memory to be uninitialized"
1094 depends on EMBEDDED && !MMU
1097 Normally, and according to the Linux spec, anonymous memory obtained
1098 from mmap() has it's contents cleared before it is passed to
1099 userspace. Enabling this config option allows you to request that
1100 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1101 providing a huge performance boost. If this option is not enabled,
1102 then the flag will be ignored.
1104 This is taken advantage of by uClibc's malloc(), and also by
1105 ELF-FDPIC binfmt's brk and stack allocator.
1107 Because of the obvious security issues, this option should only be
1108 enabled on embedded devices where you control what is run in
1109 userspace. Since that isn't generally a problem on no-MMU systems,
1110 it is normally safe to say Y here.
1112 See Documentation/nommu-mmap.txt for more information.
1115 bool "Profiling support (EXPERIMENTAL)"
1117 Say Y here to enable the extended profiling support mechanisms used
1118 by profilers such as OProfile.
1121 # Place an empty function call at each tracepoint site. Can be
1122 # dynamically changed for a probe function.
1127 source "arch/Kconfig"
1133 The slow work thread pool provides a number of dynamically allocated
1134 threads that can be used by the kernel to perform operations that
1135 take a relatively long time.
1137 An example of this would be CacheFiles doing a path lookup followed
1138 by a series of mkdirs and a create call, all of which have to touch
1141 See Documentation/slow-work.txt.
1143 config SLOW_WORK_DEBUG
1144 bool "Slow work debugging through debugfs"
1146 depends on SLOW_WORK && DEBUG_FS
1148 Display the contents of the slow work run queue through debugfs,
1149 including items currently executing.
1151 See Documentation/slow-work.txt.
1153 endmenu # General setup
1155 config HAVE_GENERIC_DMA_COHERENT
1162 depends on SLAB || SLUB_DEBUG
1170 default 0 if BASE_FULL
1171 default 1 if !BASE_FULL
1174 bool "Enable loadable module support"
1176 Kernel modules are small pieces of compiled code which can
1177 be inserted in the running kernel, rather than being
1178 permanently built into the kernel. You use the "modprobe"
1179 tool to add (and sometimes remove) them. If you say Y here,
1180 many parts of the kernel can be built as modules (by
1181 answering M instead of Y where indicated): this is most
1182 useful for infrequently used options which are not required
1183 for booting. For more information, see the man pages for
1184 modprobe, lsmod, modinfo, insmod and rmmod.
1186 If you say Y here, you will need to run "make
1187 modules_install" to put the modules under /lib/modules/
1188 where modprobe can find them (you may need to be root to do
1195 config MODULE_FORCE_LOAD
1196 bool "Forced module loading"
1199 Allow loading of modules without version information (ie. modprobe
1200 --force). Forced module loading sets the 'F' (forced) taint flag and
1201 is usually a really bad idea.
1203 config MODULE_UNLOAD
1204 bool "Module unloading"
1206 Without this option you will not be able to unload any
1207 modules (note that some modules may not be unloadable
1208 anyway), which makes your kernel smaller, faster
1209 and simpler. If unsure, say Y.
1211 config MODULE_FORCE_UNLOAD
1212 bool "Forced module unloading"
1213 depends on MODULE_UNLOAD && EXPERIMENTAL
1215 This option allows you to force a module to unload, even if the
1216 kernel believes it is unsafe: the kernel will remove the module
1217 without waiting for anyone to stop using it (using the -f option to
1218 rmmod). This is mainly for kernel developers and desperate users.
1222 bool "Module versioning support"
1224 Usually, you have to use modules compiled with your kernel.
1225 Saying Y here makes it sometimes possible to use modules
1226 compiled for different kernels, by adding enough information
1227 to the modules to (hopefully) spot any changes which would
1228 make them incompatible with the kernel you are running. If
1231 config MODULE_SRCVERSION_ALL
1232 bool "Source checksum for all modules"
1234 Modules which contain a MODULE_VERSION get an extra "srcversion"
1235 field inserted into their modinfo section, which contains a
1236 sum of the source files which made it. This helps maintainers
1237 see exactly which source was used to build a module (since
1238 others sometimes change the module source without updating
1239 the version). With this option, such a "srcversion" field
1240 will be created for all modules. If unsure, say N.
1244 config INIT_ALL_POSSIBLE
1247 Back when each arch used to define their own cpu_online_map and
1248 cpu_possible_map, some of them chose to initialize cpu_possible_map
1249 with all 1s, and others with all 0s. When they were centralised,
1250 it was better to provide this option than to break all the archs
1251 and have several arch maintainers pursuing me down dark alleys.
1256 depends on (SMP && MODULE_UNLOAD) || HOTPLUG_CPU
1258 Need stop_machine() primitive.
1260 source "block/Kconfig"
1262 config PREEMPT_NOTIFIERS
1265 source "kernel/Kconfig.locks"