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"
22 bool "Prompt for development and/or incomplete code/drivers"
24 Some of the various things that Linux supports (such as network
25 drivers, file systems, network protocols, etc.) can be in a state
26 of development where the functionality, stability, or the level of
27 testing is not yet high enough for general use. This is usually
28 known as the "alpha-test" phase among developers. If a feature is
29 currently in alpha-test, then the developers usually discourage
30 uninformed widespread use of this feature by the general public to
31 avoid "Why doesn't this work?" type mail messages. However, active
32 testing and use of these systems is welcomed. Just be aware that it
33 may not meet the normal level of reliability or it may fail to work
34 in some special cases. Detailed bug reports from people familiar
35 with the kernel internals are usually welcomed by the developers
36 (before submitting bug reports, please read the documents
37 <file:README>, <file:MAINTAINERS>, <file:REPORTING-BUGS>,
38 <file:Documentation/BUG-HUNTING>, and
39 <file:Documentation/oops-tracing.txt> in the kernel source).
41 This option will also make obsoleted drivers available. These are
42 drivers that have been replaced by something else, and/or are
43 scheduled to be removed in a future kernel release.
45 Unless you intend to help test and develop a feature or driver that
46 falls into this category, or you have a situation that requires
47 using these features, you should probably say N here, which will
48 cause the configurator to present you with fewer choices. If
49 you say Y here, you will be offered the choice of using features or
50 drivers that are currently considered to be in the alpha-test phase.
57 depends on BROKEN || !SMP
62 depends on SMP || PREEMPT
65 config INIT_ENV_ARG_LIMIT
70 Maximum of each of the number of arguments and environment
71 variables passed to init from the kernel command line.
75 string "Local version - append to kernel release"
77 Append an extra string to the end of your kernel version.
78 This will show up when you type uname, for example.
79 The string you set here will be appended after the contents of
80 any files with a filename matching localversion* in your
81 object and source tree, in that order. Your total string can
82 be a maximum of 64 characters.
84 config LOCALVERSION_AUTO
85 bool "Automatically append version information to the version string"
88 This will try to automatically determine if the current tree is a
89 release tree by looking for git tags that belong to the current
92 A string of the format -gxxxxxxxx will be added to the localversion
93 if a git-based tree is found. The string generated by this will be
94 appended after any matching localversion* files, and after the value
95 set in CONFIG_LOCALVERSION.
97 (The actual string used here is the first eight characters produced
98 by running the command:
100 $ git rev-parse --verify HEAD
102 which is done within the script "scripts/setlocalversion".)
105 bool "Support for paging of anonymous memory (swap)"
106 depends on MMU && BLOCK
109 This option allows you to choose whether you want to have support
110 for so called swap devices or swap files in your kernel that are
111 used to provide more virtual memory than the actual RAM present
112 in your computer. If unsure say Y.
117 Inter Process Communication is a suite of library functions and
118 system calls which let processes (running programs) synchronize and
119 exchange information. It is generally considered to be a good thing,
120 and some programs won't run unless you say Y here. In particular, if
121 you want to run the DOS emulator dosemu under Linux (read the
122 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
123 you'll need to say Y here.
125 You can find documentation about IPC with "info ipc" and also in
126 section 6.4 of the Linux Programmer's Guide, available from
127 <http://www.tldp.org/guides.html>.
129 config SYSVIPC_SYSCTL
136 bool "POSIX Message Queues"
137 depends on NET && EXPERIMENTAL
139 POSIX variant of message queues is a part of IPC. In POSIX message
140 queues every message has a priority which decides about succession
141 of receiving it by a process. If you want to compile and run
142 programs written e.g. for Solaris with use of its POSIX message
143 queues (functions mq_*) say Y here.
145 POSIX message queues are visible as a filesystem called 'mqueue'
146 and can be mounted somewhere if you want to do filesystem
147 operations on message queues.
151 config BSD_PROCESS_ACCT
152 bool "BSD Process Accounting"
154 If you say Y here, a user level program will be able to instruct the
155 kernel (via a special system call) to write process accounting
156 information to a file: whenever a process exits, information about
157 that process will be appended to the file by the kernel. The
158 information includes things such as creation time, owning user,
159 command name, memory usage, controlling terminal etc. (the complete
160 list is in the struct acct in <file:include/linux/acct.h>). It is
161 up to the user level program to do useful things with this
162 information. This is generally a good idea, so say Y.
164 config BSD_PROCESS_ACCT_V3
165 bool "BSD Process Accounting version 3 file format"
166 depends on BSD_PROCESS_ACCT
169 If you say Y here, the process accounting information is written
170 in a new file format that also logs the process IDs of each
171 process and it's parent. Note that this file format is incompatible
172 with previous v0/v1/v2 file formats, so you will need updated tools
173 for processing it. A preliminary version of these tools is available
174 at <http://www.gnu.org/software/acct/>.
177 bool "Export task/process statistics through netlink (EXPERIMENTAL)"
181 Export selected statistics for tasks/processes through the
182 generic netlink interface. Unlike BSD process accounting, the
183 statistics are available during the lifetime of tasks/processes as
184 responses to commands. Like BSD accounting, they are sent to user
189 config TASK_DELAY_ACCT
190 bool "Enable per-task delay accounting (EXPERIMENTAL)"
193 Collect information on time spent by a task waiting for system
194 resources like cpu, synchronous block I/O completion and swapping
195 in pages. Such statistics can help in setting a task's priorities
196 relative to other tasks for cpu, io, rss limits etc.
201 bool "Enable extended accounting over taskstats (EXPERIMENTAL)"
204 Collect extended task accounting data and send the data
205 to userland for processing over the taskstats interface.
209 config TASK_IO_ACCOUNTING
210 bool "Enable per-task storage I/O accounting (EXPERIMENTAL)"
211 depends on TASK_XACCT
213 Collect information on the number of bytes of storage I/O which this
219 bool "Auditing support"
222 Enable auditing infrastructure that can be used with another
223 kernel subsystem, such as SELinux (which requires this for
224 logging of avc messages output). Does not do system-call
225 auditing without CONFIG_AUDITSYSCALL.
228 bool "Enable system-call auditing support"
229 depends on AUDIT && (X86 || PPC || PPC64 || S390 || IA64 || UML || SPARC64|| SUPERH)
230 default y if SECURITY_SELINUX
232 Enable low-overhead system-call auditing infrastructure that
233 can be used independently or with another kernel subsystem,
234 such as SELinux. To use audit's filesystem watch feature, please
235 ensure that INOTIFY is configured.
239 depends on AUDITSYSCALL && INOTIFY
242 tristate "Kernel .config support"
244 This option enables the complete Linux kernel ".config" file
245 contents to be saved in the kernel. It provides documentation
246 of which kernel options are used in a running kernel or in an
247 on-disk kernel. This information can be extracted from the kernel
248 image file with the script scripts/extract-ikconfig and used as
249 input to rebuild the current kernel or to build another kernel.
250 It can also be extracted from a running kernel by reading
251 /proc/config.gz if enabled (below).
254 bool "Enable access to .config through /proc/config.gz"
255 depends on IKCONFIG && PROC_FS
257 This option enables access to the kernel configuration file
258 through /proc/config.gz.
261 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
265 Select kernel log buffer size as a power of 2.
275 bool "Control Group support"
277 This option will let you use process cgroup subsystems
283 bool "Example debug cgroup subsystem"
287 This option enables a simple cgroup subsystem that
288 exports useful debugging information about the cgroups
294 bool "Namespace cgroup subsystem"
297 Provides a simple namespace cgroup subsystem to
298 provide hierarchical naming of sets of namespaces,
299 for instance virtual servers and checkpoint/restart
302 config CGROUP_FREEZER
303 bool "control group freezer subsystem"
306 Provides a way to freeze and unfreeze all tasks in a
310 bool "Device controller for cgroups"
311 depends on CGROUPS && EXPERIMENTAL
313 Provides a cgroup implementing whitelists for devices which
314 a process in the cgroup can mknod or open.
317 bool "Cpuset support"
318 depends on SMP && CGROUPS
320 This option will let you create and manage CPUSETs which
321 allow dynamically partitioning a system into sets of CPUs and
322 Memory Nodes and assigning tasks to run only within those sets.
323 This is primarily useful on large SMP or NUMA systems.
328 # Architectures with an unreliable sched_clock() should select this:
330 config HAVE_UNSTABLE_SCHED_CLOCK
334 bool "Group CPU scheduler"
335 depends on EXPERIMENTAL
338 This feature lets CPU scheduler recognize task groups and control CPU
339 bandwidth allocation to such task groups.
341 config FAIR_GROUP_SCHED
342 bool "Group scheduling for SCHED_OTHER"
343 depends on GROUP_SCHED
346 config RT_GROUP_SCHED
347 bool "Group scheduling for SCHED_RR/FIFO"
348 depends on EXPERIMENTAL
349 depends on GROUP_SCHED
352 This feature lets you explicitly allocate real CPU bandwidth
353 to users or control groups (depending on the "Basis for grouping tasks"
354 setting below. If enabled, it will also make it impossible to
355 schedule realtime tasks for non-root users until you allocate
356 realtime bandwidth for them.
357 See Documentation/scheduler/sched-rt-group.txt for more information.
360 depends on GROUP_SCHED
361 prompt "Basis for grouping tasks"
367 This option will choose userid as the basis for grouping
368 tasks, thus providing equal CPU bandwidth to each user.
371 bool "Control groups"
374 This option allows you to create arbitrary task groups
375 using the "cgroup" pseudo filesystem and control
376 the cpu bandwidth allocated to each such task group.
377 Refer to Documentation/cgroups.txt for more information
378 on "cgroup" pseudo filesystem.
382 config CGROUP_CPUACCT
383 bool "Simple CPU accounting cgroup subsystem"
386 Provides a simple Resource Controller for monitoring the
387 total CPU consumed by the tasks in a cgroup
389 config RESOURCE_COUNTERS
390 bool "Resource counters"
392 This option enables controller independent resource accounting
393 infrastructure that works with cgroups
399 config CGROUP_MEM_RES_CTLR
400 bool "Memory Resource Controller for Control Groups"
401 depends on CGROUPS && RESOURCE_COUNTERS
404 Provides a memory resource controller that manages both anonymous
405 memory and page cache. (See Documentation/controllers/memory.txt)
407 Note that setting this option increases fixed memory overhead
408 associated with each page of memory in the system. By this,
409 20(40)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory
410 usage tracking struct at boot. Total amount of this is printed out
413 Only enable when you're ok with these trade offs and really
414 sure you need the memory resource controller. Even when you enable
415 this, you can set "cgroup_disable=memory" at your boot option to
416 disable memory resource controller and you can avoid overheads.
417 (and lose benefits of memory resource contoller)
419 This config option also selects MM_OWNER config option, which
420 could in turn add some fork/exit overhead.
422 config SYSFS_DEPRECATED
425 config SYSFS_DEPRECATED_V2
426 bool "Create deprecated sysfs layout for older userspace tools"
429 select SYSFS_DEPRECATED
431 This option switches the layout of sysfs to the deprecated
434 The current sysfs layout features a unified device tree at
435 /sys/devices/, which is able to express a hierarchy between
436 class devices. If the deprecated option is set to Y, the
437 unified device tree is split into a bus device tree at
438 /sys/devices/ and several individual class device trees at
439 /sys/class/. The class and bus devices will be connected by
440 "<subsystem>:<name>" and the "device" links. The "block"
441 class devices, will not show up in /sys/class/block/. Some
442 subsystems will suppress the creation of some devices which
443 depend on the unified device tree.
445 This option is not a pure compatibility option that can
446 be safely enabled on newer distributions. It will change the
447 layout of sysfs to the non-extensible deprecated version,
448 and disable some features, which can not be exported without
449 confusing older userspace tools. Since 2007/2008 all major
450 distributions do not enable this option, and ship no tools which
451 depend on the deprecated layout or this option.
453 If you are using a new kernel on an older distribution, or use
454 older userspace tools, you might need to say Y here. Do not say Y,
455 if the original kernel, that came with your distribution, has
456 this option set to N.
458 config PROC_PID_CPUSET
459 bool "Include legacy /proc/<pid>/cpuset file"
464 bool "Kernel->user space relay support (formerly relayfs)"
466 This option enables support for relay interface support in
467 certain file systems (such as debugfs).
468 It is designed to provide an efficient mechanism for tools and
469 facilities to relay large amounts of data from kernel space to
475 bool "Namespaces support" if EMBEDDED
478 Provides the way to make tasks work with different objects using
479 the same id. For example same IPC id may refer to different objects
480 or same user id or pid may refer to different tasks when used in
481 different namespaces.
485 depends on NAMESPACES
487 In this namespace tasks see different info provided with the
492 depends on NAMESPACES && SYSVIPC
494 In this namespace tasks work with IPC ids which correspond to
495 different IPC objects in different namespaces
498 bool "User namespace (EXPERIMENTAL)"
499 depends on NAMESPACES && EXPERIMENTAL
501 This allows containers, i.e. vservers, to use user namespaces
502 to provide different user info for different servers.
506 bool "PID Namespaces (EXPERIMENTAL)"
508 depends on NAMESPACES && EXPERIMENTAL
510 Support process id namespaces. This allows having multiple
511 process with the same pid as long as they are in different
512 pid namespaces. This is a building block of containers.
514 Unless you want to work with an experimental feature
517 config BLK_DEV_INITRD
518 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
519 depends on BROKEN || !FRV
521 The initial RAM filesystem is a ramfs which is loaded by the
522 boot loader (loadlin or lilo) and that is mounted as root
523 before the normal boot procedure. It is typically used to
524 load modules needed to mount the "real" root file system,
525 etc. See <file:Documentation/initrd.txt> for details.
527 If RAM disk support (BLK_DEV_RAM) is also included, this
528 also enables initial RAM disk (initrd) support and adds
529 15 Kbytes (more on some other architectures) to the kernel size.
539 config CC_OPTIMIZE_FOR_SIZE
540 bool "Optimize for size"
543 Enabling this option will pass "-Os" instead of "-O2" to gcc
544 resulting in a smaller kernel.
552 bool "Configure standard kernel features (for small systems)"
554 This option allows certain base kernel options and settings
555 to be disabled or tweaked. This is for specialized
556 environments which can tolerate a "non-standard" kernel.
557 Only use this if you really know what you are doing.
560 bool "Enable 16-bit UID system calls" if EMBEDDED
561 depends on ARM || BLACKFIN || CRIS || FRV || H8300 || X86_32 || M68K || (S390 && !64BIT) || SUPERH || SPARC32 || (SPARC64 && COMPAT) || UML || (X86_64 && IA32_EMULATION)
564 This enables the legacy 16-bit UID syscall wrappers.
566 config SYSCTL_SYSCALL
567 bool "Sysctl syscall support" if EMBEDDED
571 sys_sysctl uses binary paths that have been found challenging
572 to properly maintain and use. The interface in /proc/sys
573 using paths with ascii names is now the primary path to this
576 Almost nothing using the binary sysctl interface so if you are
577 trying to save some space it is probably safe to disable this,
578 making your kernel marginally smaller.
580 If unsure say Y here.
583 bool "Load all symbols for debugging/ksymoops" if EMBEDDED
586 Say Y here to let the kernel print out symbolic crash information and
587 symbolic stack backtraces. This increases the size of the kernel
588 somewhat, as all symbols have to be loaded into the kernel image.
591 bool "Include all symbols in kallsyms"
592 depends on DEBUG_KERNEL && KALLSYMS
594 Normally kallsyms only contains the symbols of functions, for nicer
595 OOPS messages. Some debuggers can use kallsyms for other
596 symbols too: say Y here to include all symbols, if you need them
597 and you don't care about adding 300k to the size of your kernel.
601 config KALLSYMS_STRIP_GENERATED
602 bool "Strip machine generated symbols from kallsyms"
603 depends on KALLSYMS_ALL
606 Say N if you want kallsyms to retain even machine generated symbols.
608 config KALLSYMS_EXTRA_PASS
609 bool "Do an extra kallsyms pass"
612 If kallsyms is not working correctly, the build will fail with
613 inconsistent kallsyms data. If that occurs, log a bug report and
614 turn on KALLSYMS_EXTRA_PASS which should result in a stable build.
615 Always say N here unless you find a bug in kallsyms, which must be
616 reported. KALLSYMS_EXTRA_PASS is only a temporary workaround while
617 you wait for kallsyms to be fixed.
621 bool "Support for hot-pluggable devices" if EMBEDDED
624 This option is provided for the case where no hotplug or uevent
625 capabilities is wanted by the kernel. You should only consider
626 disabling this option for embedded systems that do not use modules, a
627 dynamic /dev tree, or dynamic device discovery. Just say Y.
631 bool "Enable support for printk" if EMBEDDED
633 This option enables normal printk support. Removing it
634 eliminates most of the message strings from the kernel image
635 and makes the kernel more or less silent. As this makes it
636 very difficult to diagnose system problems, saying N here is
637 strongly discouraged.
640 bool "BUG() support" if EMBEDDED
643 Disabling this option eliminates support for BUG and WARN, reducing
644 the size of your kernel image and potentially quietly ignoring
645 numerous fatal conditions. You should only consider disabling this
646 option for embedded systems with no facilities for reporting errors.
651 bool "Enable ELF core dumps" if EMBEDDED
653 Enable support for generating core dumps. Disabling saves about 4k.
655 config PCSPKR_PLATFORM
656 bool "Enable PC-Speaker support" if EMBEDDED
657 depends on ALPHA || X86 || MIPS || PPC_PREP || PPC_CHRP || PPC_PSERIES
660 This option allows to disable the internal PC-Speaker
661 support, saving some memory.
664 bool "Disable heap randomization"
667 Randomizing heap placement makes heap exploits harder, but it
668 also breaks ancient binaries (including anything libc5 based).
669 This option changes the bootup default to heap randomization
670 disabled, and can be overriden runtime by setting
671 /proc/sys/kernel/randomize_va_space to 2.
673 On non-ancient distros (post-2000 ones) N is usually a safe choice.
677 bool "Enable full-sized data structures for core" if EMBEDDED
679 Disabling this option reduces the size of miscellaneous core
680 kernel data structures. This saves memory on small machines,
681 but may reduce performance.
684 bool "Enable futex support" if EMBEDDED
688 Disabling this option will cause the kernel to be built without
689 support for "fast userspace mutexes". The resulting kernel may not
690 run glibc-based applications correctly.
696 bool "Enable eventpoll support" if EMBEDDED
700 Disabling this option will cause the kernel to be built without
701 support for epoll family of system calls.
704 bool "Enable signalfd() system call" if EMBEDDED
708 Enable the signalfd() system call that allows to receive signals
709 on a file descriptor.
714 bool "Enable timerfd() system call" if EMBEDDED
718 Enable the timerfd() system call that allows to receive timer
719 events on a file descriptor.
724 bool "Enable eventfd() system call" if EMBEDDED
728 Enable the eventfd() system call that allows to receive both
729 kernel notification (ie. KAIO) or userspace notifications.
734 bool "Use full shmem filesystem" if EMBEDDED
738 The shmem is an internal filesystem used to manage shared memory.
739 It is backed by swap and manages resource limits. It is also exported
740 to userspace as tmpfs if TMPFS is enabled. Disabling this
741 option replaces shmem and tmpfs with the much simpler ramfs code,
742 which may be appropriate on small systems without swap.
745 bool "Enable AIO support" if EMBEDDED
748 This option enables POSIX asynchronous I/O which may by used
749 by some high performance threaded applications. Disabling
750 this option saves about 7k.
752 config VM_EVENT_COUNTERS
754 bool "Enable VM event counters for /proc/vmstat" if EMBEDDED
756 VM event counters are needed for event counts to be shown.
757 This option allows the disabling of the VM event counters
758 on EMBEDDED systems. /proc/vmstat will only show page counts
759 if VM event counters are disabled.
763 bool "Enable PCI quirk workarounds" if EMBEDDED
766 This enables workarounds for various PCI chipset
767 bugs/quirks. Disable this only if your target machine is
768 unaffected by PCI quirks.
772 bool "Enable SLUB debugging support" if EMBEDDED
773 depends on SLUB && SYSFS
775 SLUB has extensive debug support features. Disabling these can
776 result in significant savings in code size. This also disables
777 SLUB sysfs support. /sys/slab will not exist and there will be
778 no support for cache validation etc.
781 prompt "Choose SLAB allocator"
784 This option allows to select a slab allocator.
789 The regular slab allocator that is established and known to work
790 well in all environments. It organizes cache hot objects in
791 per cpu and per node queues.
794 bool "SLUB (Unqueued Allocator)"
796 SLUB is a slab allocator that minimizes cache line usage
797 instead of managing queues of cached objects (SLAB approach).
798 Per cpu caching is realized using slabs of objects instead
799 of queues of objects. SLUB can use memory efficiently
800 and has enhanced diagnostics. SLUB is the default choice for
805 bool "SLOB (Simple Allocator)"
807 SLOB replaces the stock allocator with a drastically simpler
808 allocator. SLOB is generally more space efficient but
809 does not perform as well on large systems.
814 bool "Profiling support (EXPERIMENTAL)"
816 Say Y here to enable the extended profiling support mechanisms used
817 by profilers such as OProfile.
820 # Place an empty function call at each tracepoint site. Can be
821 # dynamically changed for a probe function.
827 bool "Activate markers"
828 depends on TRACEPOINTS
830 Place an empty function call at each marker site. Can be
831 dynamically changed for a probe function.
833 source "arch/Kconfig"
835 endmenu # General setup
837 config HAVE_GENERIC_DMA_COHERENT
844 depends on SLAB || SLUB_DEBUG
853 default 0 if BASE_FULL
854 default 1 if !BASE_FULL
857 bool "Enable loadable module support"
859 Kernel modules are small pieces of compiled code which can
860 be inserted in the running kernel, rather than being
861 permanently built into the kernel. You use the "modprobe"
862 tool to add (and sometimes remove) them. If you say Y here,
863 many parts of the kernel can be built as modules (by
864 answering M instead of Y where indicated): this is most
865 useful for infrequently used options which are not required
866 for booting. For more information, see the man pages for
867 modprobe, lsmod, modinfo, insmod and rmmod.
869 If you say Y here, you will need to run "make
870 modules_install" to put the modules under /lib/modules/
871 where modprobe can find them (you may need to be root to do
878 config MODULE_FORCE_LOAD
879 bool "Forced module loading"
882 Allow loading of modules without version information (ie. modprobe
883 --force). Forced module loading sets the 'F' (forced) taint flag and
884 is usually a really bad idea.
887 bool "Module unloading"
889 Without this option you will not be able to unload any
890 modules (note that some modules may not be unloadable
891 anyway), which makes your kernel smaller, faster
892 and simpler. If unsure, say Y.
894 config MODULE_FORCE_UNLOAD
895 bool "Forced module unloading"
896 depends on MODULE_UNLOAD && EXPERIMENTAL
898 This option allows you to force a module to unload, even if the
899 kernel believes it is unsafe: the kernel will remove the module
900 without waiting for anyone to stop using it (using the -f option to
901 rmmod). This is mainly for kernel developers and desperate users.
905 bool "Module versioning support"
907 Usually, you have to use modules compiled with your kernel.
908 Saying Y here makes it sometimes possible to use modules
909 compiled for different kernels, by adding enough information
910 to the modules to (hopefully) spot any changes which would
911 make them incompatible with the kernel you are running. If
914 config MODULE_SRCVERSION_ALL
915 bool "Source checksum for all modules"
917 Modules which contain a MODULE_VERSION get an extra "srcversion"
918 field inserted into their modinfo section, which contains a
919 sum of the source files which made it. This helps maintainers
920 see exactly which source was used to build a module (since
921 others sometimes change the module source without updating
922 the version). With this option, such a "srcversion" field
923 will be created for all modules. If unsure, say N.
927 config INIT_ALL_POSSIBLE
930 Back when each arch used to define their own cpu_online_map and
931 cpu_possible_map, some of them chose to initialize cpu_possible_map
932 with all 1s, and others with all 0s. When they were centralised,
933 it was better to provide this option than to break all the archs
934 and have several arch maintainers persuing me down dark alleys.
939 depends on (SMP && MODULE_UNLOAD) || HOTPLUG_CPU
941 Need stop_machine() primitive.
943 source "block/Kconfig"
945 config PREEMPT_NOTIFIERS
949 prompt "RCU Implementation"
955 This option selects the classic RCU implementation that is
956 designed for best read-side performance on non-realtime
959 Select this option if you are unsure.
962 bool "Tree-based hierarchical RCU"
964 This option selects the RCU implementation that is
965 designed for very large SMP system with hundreds or
969 bool "Preemptible RCU"
972 This option reduces the latency of the kernel by making certain
973 RCU sections preemptible. Normally RCU code is non-preemptible, if
974 this option is selected then read-only RCU sections become
975 preemptible. This helps latency, but may expose bugs due to
976 now-naive assumptions about each RCU read-side critical section
977 remaining on a given CPU through its execution.
982 bool "Enable tracing for RCU"
983 depends on TREE_RCU || PREEMPT_RCU
985 This option provides tracing in RCU which presents stats
986 in debugfs for debugging RCU implementation.
988 Say Y here if you want to enable RCU tracing
989 Say N if you are unsure.
992 int "Tree-based hierarchical RCU fanout value"
999 This option controls the fanout of hierarchical implementations
1000 of RCU, allowing RCU to work efficiently on machines with
1001 large numbers of CPUs. This value must be at least the cube
1002 root of NR_CPUS, which allows NR_CPUS up to 32,768 for 32-bit
1003 systems and up to 262,144 for 64-bit systems.
1005 Select a specific number if testing RCU itself.
1006 Take the default if unsure.
1008 config RCU_FANOUT_EXACT
1009 bool "Disable tree-based hierarchical RCU auto-balancing"
1013 This option forces use of the exact RCU_FANOUT value specified,
1014 regardless of imbalances in the hierarchy. This is useful for
1015 testing RCU itself, and might one day be useful on systems with
1016 strong NUMA behavior.
1018 Without RCU_FANOUT_EXACT, the code will balance the hierarchy.
1022 config TREE_RCU_TRACE
1023 def_bool RCU_TRACE && TREE_RCU
1026 This option provides tracing for the TREE_RCU implementation,
1027 permitting Makefile to trivially select kernel/rcutree_trace.c.
1029 config PREEMPT_RCU_TRACE
1030 def_bool RCU_TRACE && PREEMPT_RCU
1033 This option provides tracing for the PREEMPT_RCU implementation,
1034 permitting Makefile to trivially select kernel/rcupreempt_trace.c.
projects / linux-2.6 / blob