blob: d505d9067c898409487739c831cfcff50a1a4693
1 | config ARCH |
2 | string |
3 | option env="ARCH" |
4 | |
5 | config KERNELVERSION |
6 | string |
7 | option env="KERNELVERSION" |
8 | |
9 | config DEFCONFIG_LIST |
10 | string |
11 | depends on !UML |
12 | option defconfig_list |
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" |
18 | |
19 | config CONSTRUCTORS |
20 | bool |
21 | depends on !UML |
22 | |
23 | config IRQ_WORK |
24 | bool |
25 | |
26 | config BUILDTIME_EXTABLE_SORT |
27 | bool |
28 | |
29 | config THREAD_INFO_IN_TASK |
30 | bool |
31 | help |
32 | Select this to move thread_info off the stack into task_struct. To |
33 | make this work, an arch will need to remove all thread_info fields |
34 | except flags and fix any runtime bugs. |
35 | |
36 | One subtle change that will be needed is to use try_get_task_stack() |
37 | and put_task_stack() in save_thread_stack_tsk() and get_wchan(). |
38 | |
39 | menu "General setup" |
40 | |
41 | config BROKEN |
42 | bool |
43 | |
44 | config BROKEN_ON_SMP |
45 | bool |
46 | depends on BROKEN || !SMP |
47 | default y |
48 | |
49 | config INIT_ENV_ARG_LIMIT |
50 | int |
51 | default 32 if !UML |
52 | default 128 if UML |
53 | help |
54 | Maximum of each of the number of arguments and environment |
55 | variables passed to init from the kernel command line. |
56 | |
57 | |
58 | config CROSS_COMPILE |
59 | string "Cross-compiler tool prefix" |
60 | help |
61 | Same as running 'make CROSS_COMPILE=prefix-' but stored for |
62 | default make runs in this kernel build directory. You don't |
63 | need to set this unless you want the configured kernel build |
64 | directory to select the cross-compiler automatically. |
65 | |
66 | config COMPILE_TEST |
67 | bool "Compile also drivers which will not load" |
68 | depends on !UML |
69 | default n |
70 | help |
71 | Some drivers can be compiled on a different platform than they are |
72 | intended to be run on. Despite they cannot be loaded there (or even |
73 | when they load they cannot be used due to missing HW support), |
74 | developers still, opposing to distributors, might want to build such |
75 | drivers to compile-test them. |
76 | |
77 | If you are a developer and want to build everything available, say Y |
78 | here. If you are a user/distributor, say N here to exclude useless |
79 | drivers to be distributed. |
80 | |
81 | config LOCALVERSION |
82 | string "Local version - append to kernel release" |
83 | help |
84 | Append an extra string to the end of your kernel version. |
85 | This will show up when you type uname, for example. |
86 | The string you set here will be appended after the contents of |
87 | any files with a filename matching localversion* in your |
88 | object and source tree, in that order. Your total string can |
89 | be a maximum of 64 characters. |
90 | |
91 | config LOCALVERSION_AUTO |
92 | bool "Automatically append version information to the version string" |
93 | default y |
94 | depends on !COMPILE_TEST |
95 | help |
96 | This will try to automatically determine if the current tree is a |
97 | release tree by looking for git tags that belong to the current |
98 | top of tree revision. |
99 | |
100 | A string of the format -gxxxxxxxx will be added to the localversion |
101 | if a git-based tree is found. The string generated by this will be |
102 | appended after any matching localversion* files, and after the value |
103 | set in CONFIG_LOCALVERSION. |
104 | |
105 | (The actual string used here is the first eight characters produced |
106 | by running the command: |
107 | |
108 | $ git rev-parse --verify HEAD |
109 | |
110 | which is done within the script "scripts/setlocalversion".) |
111 | |
112 | config HAVE_KERNEL_GZIP |
113 | bool |
114 | |
115 | config HAVE_KERNEL_BZIP2 |
116 | bool |
117 | |
118 | config HAVE_KERNEL_LZMA |
119 | bool |
120 | |
121 | config HAVE_KERNEL_XZ |
122 | bool |
123 | |
124 | config HAVE_KERNEL_LZO |
125 | bool |
126 | |
127 | config HAVE_KERNEL_LZ4 |
128 | bool |
129 | |
130 | choice |
131 | prompt "Kernel compression mode" |
132 | default KERNEL_GZIP |
133 | depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 |
134 | help |
135 | The linux kernel is a kind of self-extracting executable. |
136 | Several compression algorithms are available, which differ |
137 | in efficiency, compression and decompression speed. |
138 | Compression speed is only relevant when building a kernel. |
139 | Decompression speed is relevant at each boot. |
140 | |
141 | If you have any problems with bzip2 or lzma compressed |
142 | kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older |
143 | version of this functionality (bzip2 only), for 2.4, was |
144 | supplied by Christian Ludwig) |
145 | |
146 | High compression options are mostly useful for users, who |
147 | are low on disk space (embedded systems), but for whom ram |
148 | size matters less. |
149 | |
150 | If in doubt, select 'gzip' |
151 | |
152 | config KERNEL_GZIP |
153 | bool "Gzip" |
154 | depends on HAVE_KERNEL_GZIP |
155 | help |
156 | The old and tried gzip compression. It provides a good balance |
157 | between compression ratio and decompression speed. |
158 | |
159 | config KERNEL_BZIP2 |
160 | bool "Bzip2" |
161 | depends on HAVE_KERNEL_BZIP2 |
162 | help |
163 | Its compression ratio and speed is intermediate. |
164 | Decompression speed is slowest among the choices. The kernel |
165 | size is about 10% smaller with bzip2, in comparison to gzip. |
166 | Bzip2 uses a large amount of memory. For modern kernels you |
167 | will need at least 8MB RAM or more for booting. |
168 | |
169 | config KERNEL_LZMA |
170 | bool "LZMA" |
171 | depends on HAVE_KERNEL_LZMA |
172 | help |
173 | This compression algorithm's ratio is best. Decompression speed |
174 | is between gzip and bzip2. Compression is slowest. |
175 | The kernel size is about 33% smaller with LZMA in comparison to gzip. |
176 | |
177 | config KERNEL_XZ |
178 | bool "XZ" |
179 | depends on HAVE_KERNEL_XZ |
180 | help |
181 | XZ uses the LZMA2 algorithm and instruction set specific |
182 | BCJ filters which can improve compression ratio of executable |
183 | code. The size of the kernel is about 30% smaller with XZ in |
184 | comparison to gzip. On architectures for which there is a BCJ |
185 | filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ |
186 | will create a few percent smaller kernel than plain LZMA. |
187 | |
188 | The speed is about the same as with LZMA: The decompression |
189 | speed of XZ is better than that of bzip2 but worse than gzip |
190 | and LZO. Compression is slow. |
191 | |
192 | config KERNEL_LZO |
193 | bool "LZO" |
194 | depends on HAVE_KERNEL_LZO |
195 | help |
196 | Its compression ratio is the poorest among the choices. The kernel |
197 | size is about 10% bigger than gzip; however its speed |
198 | (both compression and decompression) is the fastest. |
199 | |
200 | config KERNEL_LZ4 |
201 | bool "LZ4" |
202 | depends on HAVE_KERNEL_LZ4 |
203 | help |
204 | LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding. |
205 | A preliminary version of LZ4 de/compression tool is available at |
206 | <https://code.google.com/p/lz4/>. |
207 | |
208 | Its compression ratio is worse than LZO. The size of the kernel |
209 | is about 8% bigger than LZO. But the decompression speed is |
210 | faster than LZO. |
211 | |
212 | endchoice |
213 | |
214 | config DEFAULT_HOSTNAME |
215 | string "Default hostname" |
216 | default "(none)" |
217 | help |
218 | This option determines the default system hostname before userspace |
219 | calls sethostname(2). The kernel traditionally uses "(none)" here, |
220 | but you may wish to use a different default here to make a minimal |
221 | system more usable with less configuration. |
222 | |
223 | config SWAP |
224 | bool "Support for paging of anonymous memory (swap)" |
225 | depends on MMU && BLOCK |
226 | default y |
227 | help |
228 | This option allows you to choose whether you want to have support |
229 | for so called swap devices or swap files in your kernel that are |
230 | used to provide more virtual memory than the actual RAM present |
231 | in your computer. If unsure say Y. |
232 | |
233 | config SYSVIPC |
234 | bool "System V IPC" |
235 | ---help--- |
236 | Inter Process Communication is a suite of library functions and |
237 | system calls which let processes (running programs) synchronize and |
238 | exchange information. It is generally considered to be a good thing, |
239 | and some programs won't run unless you say Y here. In particular, if |
240 | you want to run the DOS emulator dosemu under Linux (read the |
241 | DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>), |
242 | you'll need to say Y here. |
243 | |
244 | You can find documentation about IPC with "info ipc" and also in |
245 | section 6.4 of the Linux Programmer's Guide, available from |
246 | <http://www.tldp.org/guides.html>. |
247 | |
248 | config SYSVIPC_SYSCTL |
249 | bool |
250 | depends on SYSVIPC |
251 | depends on SYSCTL |
252 | default y |
253 | |
254 | config POSIX_MQUEUE |
255 | bool "POSIX Message Queues" |
256 | depends on NET |
257 | ---help--- |
258 | POSIX variant of message queues is a part of IPC. In POSIX message |
259 | queues every message has a priority which decides about succession |
260 | of receiving it by a process. If you want to compile and run |
261 | programs written e.g. for Solaris with use of its POSIX message |
262 | queues (functions mq_*) say Y here. |
263 | |
264 | POSIX message queues are visible as a filesystem called 'mqueue' |
265 | and can be mounted somewhere if you want to do filesystem |
266 | operations on message queues. |
267 | |
268 | If unsure, say Y. |
269 | |
270 | config POSIX_MQUEUE_SYSCTL |
271 | bool |
272 | depends on POSIX_MQUEUE |
273 | depends on SYSCTL |
274 | default y |
275 | |
276 | config CROSS_MEMORY_ATTACH |
277 | bool "Enable process_vm_readv/writev syscalls" |
278 | depends on MMU |
279 | default y |
280 | help |
281 | Enabling this option adds the system calls process_vm_readv and |
282 | process_vm_writev which allow a process with the correct privileges |
283 | to directly read from or write to another process' address space. |
284 | See the man page for more details. |
285 | |
286 | config FHANDLE |
287 | bool "open by fhandle syscalls" if EXPERT |
288 | select EXPORTFS |
289 | default y |
290 | help |
291 | If you say Y here, a user level program will be able to map |
292 | file names to handle and then later use the handle for |
293 | different file system operations. This is useful in implementing |
294 | userspace file servers, which now track files using handles instead |
295 | of names. The handle would remain the same even if file names |
296 | get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2) |
297 | syscalls. |
298 | |
299 | config USELIB |
300 | bool "uselib syscall" |
301 | def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION |
302 | help |
303 | This option enables the uselib syscall, a system call used in the |
304 | dynamic linker from libc5 and earlier. glibc does not use this |
305 | system call. If you intend to run programs built on libc5 or |
306 | earlier, you may need to enable this syscall. Current systems |
307 | running glibc can safely disable this. |
308 | |
309 | config AUDIT |
310 | bool "Auditing support" |
311 | depends on NET |
312 | help |
313 | Enable auditing infrastructure that can be used with another |
314 | kernel subsystem, such as SELinux (which requires this for |
315 | logging of avc messages output). System call auditing is included |
316 | on architectures which support it. |
317 | |
318 | config HAVE_ARCH_AUDITSYSCALL |
319 | bool |
320 | |
321 | config AUDITSYSCALL |
322 | def_bool y |
323 | depends on AUDIT && HAVE_ARCH_AUDITSYSCALL |
324 | |
325 | config AUDIT_WATCH |
326 | def_bool y |
327 | depends on AUDITSYSCALL |
328 | select FSNOTIFY |
329 | |
330 | config AUDIT_TREE |
331 | def_bool y |
332 | depends on AUDITSYSCALL |
333 | select FSNOTIFY |
334 | |
335 | source "kernel/irq/Kconfig" |
336 | source "kernel/time/Kconfig" |
337 | |
338 | menu "CPU/Task time and stats accounting" |
339 | |
340 | config VIRT_CPU_ACCOUNTING |
341 | bool |
342 | |
343 | choice |
344 | prompt "Cputime accounting" |
345 | default TICK_CPU_ACCOUNTING if !PPC64 |
346 | default VIRT_CPU_ACCOUNTING_NATIVE if PPC64 |
347 | |
348 | # Kind of a stub config for the pure tick based cputime accounting |
349 | config TICK_CPU_ACCOUNTING |
350 | bool "Simple tick based cputime accounting" |
351 | depends on !S390 && !NO_HZ_FULL |
352 | help |
353 | This is the basic tick based cputime accounting that maintains |
354 | statistics about user, system and idle time spent on per jiffies |
355 | granularity. |
356 | |
357 | If unsure, say Y. |
358 | |
359 | config VIRT_CPU_ACCOUNTING_NATIVE |
360 | bool "Deterministic task and CPU time accounting" |
361 | depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL |
362 | select VIRT_CPU_ACCOUNTING |
363 | help |
364 | Select this option to enable more accurate task and CPU time |
365 | accounting. This is done by reading a CPU counter on each |
366 | kernel entry and exit and on transitions within the kernel |
367 | between system, softirq and hardirq state, so there is a |
368 | small performance impact. In the case of s390 or IBM POWER > 5, |
369 | this also enables accounting of stolen time on logically-partitioned |
370 | systems. |
371 | |
372 | config VIRT_CPU_ACCOUNTING_GEN |
373 | bool "Full dynticks CPU time accounting" |
374 | depends on HAVE_CONTEXT_TRACKING |
375 | depends on HAVE_VIRT_CPU_ACCOUNTING_GEN |
376 | select VIRT_CPU_ACCOUNTING |
377 | select CONTEXT_TRACKING |
378 | help |
379 | Select this option to enable task and CPU time accounting on full |
380 | dynticks systems. This accounting is implemented by watching every |
381 | kernel-user boundaries using the context tracking subsystem. |
382 | The accounting is thus performed at the expense of some significant |
383 | overhead. |
384 | |
385 | For now this is only useful if you are working on the full |
386 | dynticks subsystem development. |
387 | |
388 | If unsure, say N. |
389 | |
390 | endchoice |
391 | |
392 | config IRQ_TIME_ACCOUNTING |
393 | bool "Fine granularity task level IRQ time accounting" |
394 | depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE |
395 | help |
396 | Select this option to enable fine granularity task irq time |
397 | accounting. This is done by reading a timestamp on each |
398 | transitions between softirq and hardirq state, so there can be a |
399 | small performance impact. |
400 | |
401 | If in doubt, say N here. |
402 | |
403 | config SCHED_WALT |
404 | bool "Support window based load tracking" |
405 | depends on SMP |
406 | help |
407 | This feature will allow the scheduler to maintain a tunable window |
408 | based set of metrics for tasks and runqueues. These metrics can be |
409 | used to guide task placement as well as task frequency requirements |
410 | for cpufreq governors. |
411 | |
412 | config BSD_PROCESS_ACCT |
413 | bool "BSD Process Accounting" |
414 | depends on MULTIUSER |
415 | help |
416 | If you say Y here, a user level program will be able to instruct the |
417 | kernel (via a special system call) to write process accounting |
418 | information to a file: whenever a process exits, information about |
419 | that process will be appended to the file by the kernel. The |
420 | information includes things such as creation time, owning user, |
421 | command name, memory usage, controlling terminal etc. (the complete |
422 | list is in the struct acct in <file:include/linux/acct.h>). It is |
423 | up to the user level program to do useful things with this |
424 | information. This is generally a good idea, so say Y. |
425 | |
426 | config BSD_PROCESS_ACCT_V3 |
427 | bool "BSD Process Accounting version 3 file format" |
428 | depends on BSD_PROCESS_ACCT |
429 | default n |
430 | help |
431 | If you say Y here, the process accounting information is written |
432 | in a new file format that also logs the process IDs of each |
433 | process and it's parent. Note that this file format is incompatible |
434 | with previous v0/v1/v2 file formats, so you will need updated tools |
435 | for processing it. A preliminary version of these tools is available |
436 | at <http://www.gnu.org/software/acct/>. |
437 | |
438 | config TASKSTATS |
439 | bool "Export task/process statistics through netlink" |
440 | depends on NET |
441 | depends on MULTIUSER |
442 | default n |
443 | help |
444 | Export selected statistics for tasks/processes through the |
445 | generic netlink interface. Unlike BSD process accounting, the |
446 | statistics are available during the lifetime of tasks/processes as |
447 | responses to commands. Like BSD accounting, they are sent to user |
448 | space on task exit. |
449 | |
450 | Say N if unsure. |
451 | |
452 | config TASK_DELAY_ACCT |
453 | bool "Enable per-task delay accounting" |
454 | depends on TASKSTATS |
455 | select SCHED_INFO |
456 | help |
457 | Collect information on time spent by a task waiting for system |
458 | resources like cpu, synchronous block I/O completion and swapping |
459 | in pages. Such statistics can help in setting a task's priorities |
460 | relative to other tasks for cpu, io, rss limits etc. |
461 | |
462 | Say N if unsure. |
463 | |
464 | config TASK_XACCT |
465 | bool "Enable extended accounting over taskstats" |
466 | depends on TASKSTATS |
467 | help |
468 | Collect extended task accounting data and send the data |
469 | to userland for processing over the taskstats interface. |
470 | |
471 | Say N if unsure. |
472 | |
473 | config TASK_IO_ACCOUNTING |
474 | bool "Enable per-task storage I/O accounting" |
475 | depends on TASK_XACCT |
476 | help |
477 | Collect information on the number of bytes of storage I/O which this |
478 | task has caused. |
479 | |
480 | Say N if unsure. |
481 | |
482 | config PSI |
483 | bool "Pressure stall information tracking" |
484 | help |
485 | Collect metrics that indicate how overcommitted the CPU, memory, |
486 | and IO capacity are in the system. |
487 | |
488 | If you say Y here, the kernel will create /proc/pressure/ with the |
489 | pressure statistics files cpu, memory, and io. These will indicate |
490 | the share of walltime in which some or all tasks in the system are |
491 | delayed due to contention of the respective resource. |
492 | |
493 | In kernels with cgroup support, cgroups (cgroup2 only) will |
494 | have cpu.pressure, memory.pressure, and io.pressure files, |
495 | which aggregate pressure stalls for the grouped tasks only. |
496 | |
497 | For more details see Documentation/accounting/psi.txt. |
498 | |
499 | Say N if unsure. |
500 | |
501 | config PSI_DEFAULT_DISABLED |
502 | bool "Require boot parameter to enable pressure stall information tracking" |
503 | default n |
504 | depends on PSI |
505 | help |
506 | If set, pressure stall information tracking will be disabled |
507 | per default but can be enabled through passing psi=1 on the |
508 | kernel commandline during boot. |
509 | |
510 | This feature adds some code to the task wakeup and sleep |
511 | paths of the scheduler. The overhead is too low to affect |
512 | common scheduling-intense workloads in practice (such as |
513 | webservers, memcache), but it does show up in artificial |
514 | scheduler stress tests, such as hackbench. |
515 | |
516 | If you are paranoid and not sure what the kernel will be |
517 | used for, say Y. |
518 | |
519 | Say N if unsure. |
520 | |
521 | endmenu # "CPU/Task time and stats accounting" |
522 | |
523 | menu "RCU Subsystem" |
524 | |
525 | config TREE_RCU |
526 | bool |
527 | default y if !PREEMPT && SMP |
528 | help |
529 | This option selects the RCU implementation that is |
530 | designed for very large SMP system with hundreds or |
531 | thousands of CPUs. It also scales down nicely to |
532 | smaller systems. |
533 | |
534 | config PREEMPT_RCU |
535 | bool |
536 | default y if PREEMPT |
537 | help |
538 | This option selects the RCU implementation that is |
539 | designed for very large SMP systems with hundreds or |
540 | thousands of CPUs, but for which real-time response |
541 | is also required. It also scales down nicely to |
542 | smaller systems. |
543 | |
544 | Select this option if you are unsure. |
545 | |
546 | config TINY_RCU |
547 | bool |
548 | default y if !PREEMPT && !SMP |
549 | help |
550 | This option selects the RCU implementation that is |
551 | designed for UP systems from which real-time response |
552 | is not required. This option greatly reduces the |
553 | memory footprint of RCU. |
554 | |
555 | config RCU_EXPERT |
556 | bool "Make expert-level adjustments to RCU configuration" |
557 | default n |
558 | help |
559 | This option needs to be enabled if you wish to make |
560 | expert-level adjustments to RCU configuration. By default, |
561 | no such adjustments can be made, which has the often-beneficial |
562 | side-effect of preventing "make oldconfig" from asking you all |
563 | sorts of detailed questions about how you would like numerous |
564 | obscure RCU options to be set up. |
565 | |
566 | Say Y if you need to make expert-level adjustments to RCU. |
567 | |
568 | Say N if you are unsure. |
569 | |
570 | config SRCU |
571 | bool |
572 | help |
573 | This option selects the sleepable version of RCU. This version |
574 | permits arbitrary sleeping or blocking within RCU read-side critical |
575 | sections. |
576 | |
577 | config TASKS_RCU |
578 | bool |
579 | default n |
580 | depends on !UML |
581 | select SRCU |
582 | help |
583 | This option enables a task-based RCU implementation that uses |
584 | only voluntary context switch (not preemption!), idle, and |
585 | user-mode execution as quiescent states. |
586 | |
587 | config RCU_STALL_COMMON |
588 | def_bool ( TREE_RCU || PREEMPT_RCU || RCU_TRACE ) |
589 | help |
590 | This option enables RCU CPU stall code that is common between |
591 | the TINY and TREE variants of RCU. The purpose is to allow |
592 | the tiny variants to disable RCU CPU stall warnings, while |
593 | making these warnings mandatory for the tree variants. |
594 | |
595 | config CONTEXT_TRACKING |
596 | bool |
597 | |
598 | config CONTEXT_TRACKING_FORCE |
599 | bool "Force context tracking" |
600 | depends on CONTEXT_TRACKING |
601 | default y if !NO_HZ_FULL |
602 | help |
603 | The major pre-requirement for full dynticks to work is to |
604 | support the context tracking subsystem. But there are also |
605 | other dependencies to provide in order to make the full |
606 | dynticks working. |
607 | |
608 | This option stands for testing when an arch implements the |
609 | context tracking backend but doesn't yet fullfill all the |
610 | requirements to make the full dynticks feature working. |
611 | Without the full dynticks, there is no way to test the support |
612 | for context tracking and the subsystems that rely on it: RCU |
613 | userspace extended quiescent state and tickless cputime |
614 | accounting. This option copes with the absence of the full |
615 | dynticks subsystem by forcing the context tracking on all |
616 | CPUs in the system. |
617 | |
618 | Say Y only if you're working on the development of an |
619 | architecture backend for the context tracking. |
620 | |
621 | Say N otherwise, this option brings an overhead that you |
622 | don't want in production. |
623 | |
624 | |
625 | config RCU_FANOUT |
626 | int "Tree-based hierarchical RCU fanout value" |
627 | range 2 64 if 64BIT |
628 | range 2 32 if !64BIT |
629 | depends on (TREE_RCU || PREEMPT_RCU) && RCU_EXPERT |
630 | default 64 if 64BIT |
631 | default 32 if !64BIT |
632 | help |
633 | This option controls the fanout of hierarchical implementations |
634 | of RCU, allowing RCU to work efficiently on machines with |
635 | large numbers of CPUs. This value must be at least the fourth |
636 | root of NR_CPUS, which allows NR_CPUS to be insanely large. |
637 | The default value of RCU_FANOUT should be used for production |
638 | systems, but if you are stress-testing the RCU implementation |
639 | itself, small RCU_FANOUT values allow you to test large-system |
640 | code paths on small(er) systems. |
641 | |
642 | Select a specific number if testing RCU itself. |
643 | Take the default if unsure. |
644 | |
645 | config RCU_FANOUT_LEAF |
646 | int "Tree-based hierarchical RCU leaf-level fanout value" |
647 | range 2 64 if 64BIT |
648 | range 2 32 if !64BIT |
649 | depends on (TREE_RCU || PREEMPT_RCU) && RCU_EXPERT |
650 | default 16 |
651 | help |
652 | This option controls the leaf-level fanout of hierarchical |
653 | implementations of RCU, and allows trading off cache misses |
654 | against lock contention. Systems that synchronize their |
655 | scheduling-clock interrupts for energy-efficiency reasons will |
656 | want the default because the smaller leaf-level fanout keeps |
657 | lock contention levels acceptably low. Very large systems |
658 | (hundreds or thousands of CPUs) will instead want to set this |
659 | value to the maximum value possible in order to reduce the |
660 | number of cache misses incurred during RCU's grace-period |
661 | initialization. These systems tend to run CPU-bound, and thus |
662 | are not helped by synchronized interrupts, and thus tend to |
663 | skew them, which reduces lock contention enough that large |
664 | leaf-level fanouts work well. |
665 | |
666 | Select a specific number if testing RCU itself. |
667 | |
668 | Select the maximum permissible value for large systems. |
669 | |
670 | Take the default if unsure. |
671 | |
672 | config RCU_FAST_NO_HZ |
673 | bool "Accelerate last non-dyntick-idle CPU's grace periods" |
674 | depends on NO_HZ_COMMON && SMP && RCU_EXPERT |
675 | default n |
676 | help |
677 | This option permits CPUs to enter dynticks-idle state even if |
678 | they have RCU callbacks queued, and prevents RCU from waking |
679 | these CPUs up more than roughly once every four jiffies (by |
680 | default, you can adjust this using the rcutree.rcu_idle_gp_delay |
681 | parameter), thus improving energy efficiency. On the other |
682 | hand, this option increases the duration of RCU grace periods, |
683 | for example, slowing down synchronize_rcu(). |
684 | |
685 | Say Y if energy efficiency is critically important, and you |
686 | don't care about increased grace-period durations. |
687 | |
688 | Say N if you are unsure. |
689 | |
690 | config TREE_RCU_TRACE |
691 | def_bool RCU_TRACE && ( TREE_RCU || PREEMPT_RCU ) |
692 | select DEBUG_FS |
693 | help |
694 | This option provides tracing for the TREE_RCU and |
695 | PREEMPT_RCU implementations, permitting Makefile to |
696 | trivially select kernel/rcutree_trace.c. |
697 | |
698 | config RCU_BOOST |
699 | bool "Enable RCU priority boosting" |
700 | depends on RT_MUTEXES && PREEMPT_RCU && RCU_EXPERT |
701 | default n |
702 | help |
703 | This option boosts the priority of preempted RCU readers that |
704 | block the current preemptible RCU grace period for too long. |
705 | This option also prevents heavy loads from blocking RCU |
706 | callback invocation for all flavors of RCU. |
707 | |
708 | Say Y here if you are working with real-time apps or heavy loads |
709 | Say N here if you are unsure. |
710 | |
711 | config RCU_KTHREAD_PRIO |
712 | int "Real-time priority to use for RCU worker threads" |
713 | range 1 99 if RCU_BOOST |
714 | range 0 99 if !RCU_BOOST |
715 | default 1 if RCU_BOOST |
716 | default 0 if !RCU_BOOST |
717 | depends on RCU_EXPERT |
718 | help |
719 | This option specifies the SCHED_FIFO priority value that will be |
720 | assigned to the rcuc/n and rcub/n threads and is also the value |
721 | used for RCU_BOOST (if enabled). If you are working with a |
722 | real-time application that has one or more CPU-bound threads |
723 | running at a real-time priority level, you should set |
724 | RCU_KTHREAD_PRIO to a priority higher than the highest-priority |
725 | real-time CPU-bound application thread. The default RCU_KTHREAD_PRIO |
726 | value of 1 is appropriate in the common case, which is real-time |
727 | applications that do not have any CPU-bound threads. |
728 | |
729 | Some real-time applications might not have a single real-time |
730 | thread that saturates a given CPU, but instead might have |
731 | multiple real-time threads that, taken together, fully utilize |
732 | that CPU. In this case, you should set RCU_KTHREAD_PRIO to |
733 | a priority higher than the lowest-priority thread that is |
734 | conspiring to prevent the CPU from running any non-real-time |
735 | tasks. For example, if one thread at priority 10 and another |
736 | thread at priority 5 are between themselves fully consuming |
737 | the CPU time on a given CPU, then RCU_KTHREAD_PRIO should be |
738 | set to priority 6 or higher. |
739 | |
740 | Specify the real-time priority, or take the default if unsure. |
741 | |
742 | config RCU_BOOST_DELAY |
743 | int "Milliseconds to delay boosting after RCU grace-period start" |
744 | range 0 3000 |
745 | depends on RCU_BOOST |
746 | default 500 |
747 | help |
748 | This option specifies the time to wait after the beginning of |
749 | a given grace period before priority-boosting preempted RCU |
750 | readers blocking that grace period. Note that any RCU reader |
751 | blocking an expedited RCU grace period is boosted immediately. |
752 | |
753 | Accept the default if unsure. |
754 | |
755 | config RCU_NOCB_CPU |
756 | bool "Offload RCU callback processing from boot-selected CPUs" |
757 | depends on TREE_RCU || PREEMPT_RCU |
758 | depends on RCU_EXPERT || NO_HZ_FULL |
759 | default n |
760 | help |
761 | Use this option to reduce OS jitter for aggressive HPC or |
762 | real-time workloads. It can also be used to offload RCU |
763 | callback invocation to energy-efficient CPUs in battery-powered |
764 | asymmetric multiprocessors. |
765 | |
766 | This option offloads callback invocation from the set of |
767 | CPUs specified at boot time by the rcu_nocbs parameter. |
768 | For each such CPU, a kthread ("rcuox/N") will be created to |
769 | invoke callbacks, where the "N" is the CPU being offloaded, |
770 | and where the "x" is "b" for RCU-bh, "p" for RCU-preempt, and |
771 | "s" for RCU-sched. Nothing prevents this kthread from running |
772 | on the specified CPUs, but (1) the kthreads may be preempted |
773 | between each callback, and (2) affinity or cgroups can be used |
774 | to force the kthreads to run on whatever set of CPUs is desired. |
775 | |
776 | Say Y here if you want to help to debug reduced OS jitter. |
777 | Say N here if you are unsure. |
778 | |
779 | choice |
780 | prompt "Build-forced no-CBs CPUs" |
781 | default RCU_NOCB_CPU_NONE |
782 | depends on RCU_NOCB_CPU |
783 | help |
784 | This option allows no-CBs CPUs (whose RCU callbacks are invoked |
785 | from kthreads rather than from softirq context) to be specified |
786 | at build time. Additional no-CBs CPUs may be specified by |
787 | the rcu_nocbs= boot parameter. |
788 | |
789 | config RCU_NOCB_CPU_NONE |
790 | bool "No build_forced no-CBs CPUs" |
791 | help |
792 | This option does not force any of the CPUs to be no-CBs CPUs. |
793 | Only CPUs designated by the rcu_nocbs= boot parameter will be |
794 | no-CBs CPUs, whose RCU callbacks will be invoked by per-CPU |
795 | kthreads whose names begin with "rcuo". All other CPUs will |
796 | invoke their own RCU callbacks in softirq context. |
797 | |
798 | Select this option if you want to choose no-CBs CPUs at |
799 | boot time, for example, to allow testing of different no-CBs |
800 | configurations without having to rebuild the kernel each time. |
801 | |
802 | config RCU_NOCB_CPU_ZERO |
803 | bool "CPU 0 is a build_forced no-CBs CPU" |
804 | help |
805 | This option forces CPU 0 to be a no-CBs CPU, so that its RCU |
806 | callbacks are invoked by a per-CPU kthread whose name begins |
807 | with "rcuo". Additional CPUs may be designated as no-CBs |
808 | CPUs using the rcu_nocbs= boot parameter will be no-CBs CPUs. |
809 | All other CPUs will invoke their own RCU callbacks in softirq |
810 | context. |
811 | |
812 | Select this if CPU 0 needs to be a no-CBs CPU for real-time |
813 | or energy-efficiency reasons, but the real reason it exists |
814 | is to ensure that randconfig testing covers mixed systems. |
815 | |
816 | config RCU_NOCB_CPU_ALL |
817 | bool "All CPUs are build_forced no-CBs CPUs" |
818 | help |
819 | This option forces all CPUs to be no-CBs CPUs. The rcu_nocbs= |
820 | boot parameter will be ignored. All CPUs' RCU callbacks will |
821 | be executed in the context of per-CPU rcuo kthreads created for |
822 | this purpose. Assuming that the kthreads whose names start with |
823 | "rcuo" are bound to "housekeeping" CPUs, this reduces OS jitter |
824 | on the remaining CPUs, but might decrease memory locality during |
825 | RCU-callback invocation, thus potentially degrading throughput. |
826 | |
827 | Select this if all CPUs need to be no-CBs CPUs for real-time |
828 | or energy-efficiency reasons. |
829 | |
830 | endchoice |
831 | |
832 | config RCU_EXPEDITE_BOOT |
833 | bool |
834 | default n |
835 | help |
836 | This option enables expedited grace periods at boot time, |
837 | as if rcu_expedite_gp() had been invoked early in boot. |
838 | The corresponding rcu_unexpedite_gp() is invoked from |
839 | rcu_end_inkernel_boot(), which is intended to be invoked |
840 | at the end of the kernel-only boot sequence, just before |
841 | init is exec'ed. |
842 | |
843 | Accept the default if unsure. |
844 | |
845 | endmenu # "RCU Subsystem" |
846 | |
847 | config BUILD_BIN2C |
848 | bool |
849 | default n |
850 | |
851 | config IKCONFIG |
852 | tristate "Kernel .config support" |
853 | select BUILD_BIN2C |
854 | ---help--- |
855 | This option enables the complete Linux kernel ".config" file |
856 | contents to be saved in the kernel. It provides documentation |
857 | of which kernel options are used in a running kernel or in an |
858 | on-disk kernel. This information can be extracted from the kernel |
859 | image file with the script scripts/extract-ikconfig and used as |
860 | input to rebuild the current kernel or to build another kernel. |
861 | It can also be extracted from a running kernel by reading |
862 | /proc/config.gz if enabled (below). |
863 | |
864 | config IKCONFIG_PROC |
865 | bool "Enable access to .config through /proc/config.gz" |
866 | depends on IKCONFIG && PROC_FS |
867 | ---help--- |
868 | This option enables access to the kernel configuration file |
869 | through /proc/config.gz. |
870 | |
871 | config LOG_BUF_SHIFT |
872 | int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" |
873 | range 12 25 |
874 | default 17 |
875 | depends on PRINTK |
876 | help |
877 | Select the minimal kernel log buffer size as a power of 2. |
878 | The final size is affected by LOG_CPU_MAX_BUF_SHIFT config |
879 | parameter, see below. Any higher size also might be forced |
880 | by "log_buf_len" boot parameter. |
881 | |
882 | Examples: |
883 | 17 => 128 KB |
884 | 16 => 64 KB |
885 | 15 => 32 KB |
886 | 14 => 16 KB |
887 | 13 => 8 KB |
888 | 12 => 4 KB |
889 | |
890 | config LOG_CPU_MAX_BUF_SHIFT |
891 | int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)" |
892 | depends on SMP |
893 | range 0 21 |
894 | default 12 if !BASE_SMALL |
895 | default 0 if BASE_SMALL |
896 | depends on PRINTK |
897 | help |
898 | This option allows to increase the default ring buffer size |
899 | according to the number of CPUs. The value defines the contribution |
900 | of each CPU as a power of 2. The used space is typically only few |
901 | lines however it might be much more when problems are reported, |
902 | e.g. backtraces. |
903 | |
904 | The increased size means that a new buffer has to be allocated and |
905 | the original static one is unused. It makes sense only on systems |
906 | with more CPUs. Therefore this value is used only when the sum of |
907 | contributions is greater than the half of the default kernel ring |
908 | buffer as defined by LOG_BUF_SHIFT. The default values are set |
909 | so that more than 64 CPUs are needed to trigger the allocation. |
910 | |
911 | Also this option is ignored when "log_buf_len" kernel parameter is |
912 | used as it forces an exact (power of two) size of the ring buffer. |
913 | |
914 | The number of possible CPUs is used for this computation ignoring |
915 | hotplugging making the computation optimal for the worst case |
916 | scenario while allowing a simple algorithm to be used from bootup. |
917 | |
918 | Examples shift values and their meaning: |
919 | 17 => 128 KB for each CPU |
920 | 16 => 64 KB for each CPU |
921 | 15 => 32 KB for each CPU |
922 | 14 => 16 KB for each CPU |
923 | 13 => 8 KB for each CPU |
924 | 12 => 4 KB for each CPU |
925 | |
926 | config NMI_LOG_BUF_SHIFT |
927 | int "Temporary per-CPU NMI log buffer size (12 => 4KB, 13 => 8KB)" |
928 | range 10 21 |
929 | default 13 |
930 | depends on PRINTK_NMI |
931 | help |
932 | Select the size of a per-CPU buffer where NMI messages are temporary |
933 | stored. They are copied to the main log buffer in a safe context |
934 | to avoid a deadlock. The value defines the size as a power of 2. |
935 | |
936 | NMI messages are rare and limited. The largest one is when |
937 | a backtrace is printed. It usually fits into 4KB. Select |
938 | 8KB if you want to be on the safe side. |
939 | |
940 | Examples: |
941 | 17 => 128 KB for each CPU |
942 | 16 => 64 KB for each CPU |
943 | 15 => 32 KB for each CPU |
944 | 14 => 16 KB for each CPU |
945 | 13 => 8 KB for each CPU |
946 | 12 => 4 KB for each CPU |
947 | |
948 | # |
949 | # Architectures with an unreliable sched_clock() should select this: |
950 | # |
951 | config HAVE_UNSTABLE_SCHED_CLOCK |
952 | bool |
953 | |
954 | config GENERIC_SCHED_CLOCK |
955 | bool |
956 | |
957 | # |
958 | # For architectures that want to enable the support for NUMA-affine scheduler |
959 | # balancing logic: |
960 | # |
961 | config ARCH_SUPPORTS_NUMA_BALANCING |
962 | bool |
963 | |
964 | # |
965 | # For architectures that prefer to flush all TLBs after a number of pages |
966 | # are unmapped instead of sending one IPI per page to flush. The architecture |
967 | # must provide guarantees on what happens if a clean TLB cache entry is |
968 | # written after the unmap. Details are in mm/rmap.c near the check for |
969 | # should_defer_flush. The architecture should also consider if the full flush |
970 | # and the refill costs are offset by the savings of sending fewer IPIs. |
971 | config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH |
972 | bool |
973 | |
974 | # |
975 | # For architectures that know their GCC __int128 support is sound |
976 | # |
977 | config ARCH_SUPPORTS_INT128 |
978 | bool |
979 | |
980 | # For architectures that (ab)use NUMA to represent different memory regions |
981 | # all cpu-local but of different latencies, such as SuperH. |
982 | # |
983 | config ARCH_WANT_NUMA_VARIABLE_LOCALITY |
984 | bool |
985 | |
986 | config NUMA_BALANCING |
987 | bool "Memory placement aware NUMA scheduler" |
988 | depends on ARCH_SUPPORTS_NUMA_BALANCING |
989 | depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY |
990 | depends on SMP && NUMA && MIGRATION |
991 | help |
992 | This option adds support for automatic NUMA aware memory/task placement. |
993 | The mechanism is quite primitive and is based on migrating memory when |
994 | it has references to the node the task is running on. |
995 | |
996 | This system will be inactive on UMA systems. |
997 | |
998 | config NUMA_BALANCING_DEFAULT_ENABLED |
999 | bool "Automatically enable NUMA aware memory/task placement" |
1000 | default y |
1001 | depends on NUMA_BALANCING |
1002 | help |
1003 | If set, automatic NUMA balancing will be enabled if running on a NUMA |
1004 | machine. |
1005 | |
1006 | menuconfig CGROUPS |
1007 | bool "Control Group support" |
1008 | select KERNFS |
1009 | help |
1010 | This option adds support for grouping sets of processes together, for |
1011 | use with process control subsystems such as Cpusets, CFS, memory |
1012 | controls or device isolation. |
1013 | See |
1014 | - Documentation/scheduler/sched-design-CFS.txt (CFS) |
1015 | - Documentation/cgroup-v1/ (features for grouping, isolation |
1016 | and resource control) |
1017 | |
1018 | Say N if unsure. |
1019 | |
1020 | if CGROUPS |
1021 | |
1022 | config CGROUP_DEBUG |
1023 | bool "Example debug cgroup subsystem" |
1024 | default n |
1025 | help |
1026 | This option enables a simple cgroup subsystem that |
1027 | exports useful debugging information about the cgroups |
1028 | framework. |
1029 | |
1030 | Say N if unsure. |
1031 | |
1032 | config CGROUP_FREEZER |
1033 | bool "Freezer cgroup subsystem" |
1034 | help |
1035 | Provides a way to freeze and unfreeze all tasks in a |
1036 | cgroup. |
1037 | |
1038 | config CGROUP_PIDS |
1039 | bool "PIDs cgroup subsystem" |
1040 | help |
1041 | Provides enforcement of process number limits in the scope of a |
1042 | cgroup. Any attempt to fork more processes than is allowed in the |
1043 | cgroup will fail. PIDs are fundamentally a global resource because it |
1044 | is fairly trivial to reach PID exhaustion before you reach even a |
1045 | conservative kmemcg limit. As a result, it is possible to grind a |
1046 | system to halt without being limited by other cgroup policies. The |
1047 | PIDs cgroup subsystem is designed to stop this from happening. |
1048 | |
1049 | It should be noted that organisational operations (such as attaching |
1050 | to a cgroup hierarchy will *not* be blocked by the PIDs subsystem), |
1051 | since the PIDs limit only affects a process's ability to fork, not to |
1052 | attach to a cgroup. |
1053 | |
1054 | config CGROUP_DEVICE |
1055 | bool "Device controller for cgroups" |
1056 | help |
1057 | Provides a cgroup implementing whitelists for devices which |
1058 | a process in the cgroup can mknod or open. |
1059 | |
1060 | config CPUSETS |
1061 | bool "Cpuset support" |
1062 | help |
1063 | This option will let you create and manage CPUSETs which |
1064 | allow dynamically partitioning a system into sets of CPUs and |
1065 | Memory Nodes and assigning tasks to run only within those sets. |
1066 | This is primarily useful on large SMP or NUMA systems. |
1067 | |
1068 | Say N if unsure. |
1069 | |
1070 | config PROC_PID_CPUSET |
1071 | bool "Include legacy /proc/<pid>/cpuset file" |
1072 | depends on CPUSETS |
1073 | default y |
1074 | |
1075 | config CGROUP_CPUACCT |
1076 | bool "Simple CPU accounting cgroup subsystem" |
1077 | help |
1078 | Provides a simple Resource Controller for monitoring the |
1079 | total CPU consumed by the tasks in a cgroup. |
1080 | |
1081 | config CGROUP_SCHEDTUNE |
1082 | bool "CFS tasks boosting cgroup subsystem (EXPERIMENTAL)" |
1083 | depends on SCHED_TUNE |
1084 | help |
1085 | This option provides the "schedtune" controller which improves the |
1086 | flexibility of the task boosting mechanism by introducing the support |
1087 | to define "per task" boost values. |
1088 | |
1089 | This new controller: |
1090 | 1. allows only a two layers hierarchy, where the root defines the |
1091 | system-wide boost value and its direct childrens define each one a |
1092 | different "class of tasks" to be boosted with a different value |
1093 | 2. supports up to 16 different task classes, each one which could be |
1094 | configured with a different boost value |
1095 | |
1096 | Say N if unsure. |
1097 | |
1098 | config PAGE_COUNTER |
1099 | bool |
1100 | |
1101 | config MEMCG |
1102 | bool "Memory controller" |
1103 | select PAGE_COUNTER |
1104 | select EVENTFD |
1105 | help |
1106 | Provides control over the memory footprint of tasks in a cgroup. |
1107 | |
1108 | config MEMCG_SWAP |
1109 | bool "Swap controller" |
1110 | depends on MEMCG && SWAP |
1111 | help |
1112 | Provides control over the swap space consumed by tasks in a cgroup. |
1113 | |
1114 | config MEMCG_SWAP_ENABLED |
1115 | bool "Swap controller enabled by default" |
1116 | depends on MEMCG_SWAP |
1117 | default y |
1118 | help |
1119 | Memory Resource Controller Swap Extension comes with its price in |
1120 | a bigger memory consumption. General purpose distribution kernels |
1121 | which want to enable the feature but keep it disabled by default |
1122 | and let the user enable it by swapaccount=1 boot command line |
1123 | parameter should have this option unselected. |
1124 | For those who want to have the feature enabled by default should |
1125 | select this option (if, for some reason, they need to disable it |
1126 | then swapaccount=0 does the trick). |
1127 | |
1128 | config BLK_CGROUP |
1129 | bool "IO controller" |
1130 | depends on BLOCK |
1131 | default n |
1132 | ---help--- |
1133 | Generic block IO controller cgroup interface. This is the common |
1134 | cgroup interface which should be used by various IO controlling |
1135 | policies. |
1136 | |
1137 | Currently, CFQ IO scheduler uses it to recognize task groups and |
1138 | control disk bandwidth allocation (proportional time slice allocation) |
1139 | to such task groups. It is also used by bio throttling logic in |
1140 | block layer to implement upper limit in IO rates on a device. |
1141 | |
1142 | This option only enables generic Block IO controller infrastructure. |
1143 | One needs to also enable actual IO controlling logic/policy. For |
1144 | enabling proportional weight division of disk bandwidth in CFQ, set |
1145 | CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set |
1146 | CONFIG_BLK_DEV_THROTTLING=y. |
1147 | |
1148 | See Documentation/cgroup-v1/blkio-controller.txt for more information. |
1149 | |
1150 | config DEBUG_BLK_CGROUP |
1151 | bool "IO controller debugging" |
1152 | depends on BLK_CGROUP |
1153 | default n |
1154 | ---help--- |
1155 | Enable some debugging help. Currently it exports additional stat |
1156 | files in a cgroup which can be useful for debugging. |
1157 | |
1158 | config CGROUP_WRITEBACK |
1159 | bool |
1160 | depends on MEMCG && BLK_CGROUP |
1161 | default y |
1162 | |
1163 | menuconfig CGROUP_SCHED |
1164 | bool "CPU controller" |
1165 | default n |
1166 | help |
1167 | This feature lets CPU scheduler recognize task groups and control CPU |
1168 | bandwidth allocation to such task groups. It uses cgroups to group |
1169 | tasks. |
1170 | |
1171 | if CGROUP_SCHED |
1172 | config FAIR_GROUP_SCHED |
1173 | bool "Group scheduling for SCHED_OTHER" |
1174 | depends on CGROUP_SCHED |
1175 | default CGROUP_SCHED |
1176 | |
1177 | config CFS_BANDWIDTH |
1178 | bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED" |
1179 | depends on FAIR_GROUP_SCHED |
1180 | default n |
1181 | help |
1182 | This option allows users to define CPU bandwidth rates (limits) for |
1183 | tasks running within the fair group scheduler. Groups with no limit |
1184 | set are considered to be unconstrained and will run with no |
1185 | restriction. |
1186 | See tip/Documentation/scheduler/sched-bwc.txt for more information. |
1187 | |
1188 | config RT_GROUP_SCHED |
1189 | bool "Group scheduling for SCHED_RR/FIFO" |
1190 | depends on CGROUP_SCHED |
1191 | default n |
1192 | help |
1193 | This feature lets you explicitly allocate real CPU bandwidth |
1194 | to task groups. If enabled, it will also make it impossible to |
1195 | schedule realtime tasks for non-root users until you allocate |
1196 | realtime bandwidth for them. |
1197 | See Documentation/scheduler/sched-rt-group.txt for more information. |
1198 | |
1199 | endif #CGROUP_SCHED |
1200 | |
1201 | config CGROUP_PIDS |
1202 | bool "PIDs controller" |
1203 | help |
1204 | Provides enforcement of process number limits in the scope of a |
1205 | cgroup. Any attempt to fork more processes than is allowed in the |
1206 | cgroup will fail. PIDs are fundamentally a global resource because it |
1207 | is fairly trivial to reach PID exhaustion before you reach even a |
1208 | conservative kmemcg limit. As a result, it is possible to grind a |
1209 | system to halt without being limited by other cgroup policies. The |
1210 | PIDs controller is designed to stop this from happening. |
1211 | |
1212 | It should be noted that organisational operations (such as attaching |
1213 | to a cgroup hierarchy will *not* be blocked by the PIDs controller), |
1214 | since the PIDs limit only affects a process's ability to fork, not to |
1215 | attach to a cgroup. |
1216 | |
1217 | config CGROUP_FREEZER |
1218 | bool "Freezer controller" |
1219 | help |
1220 | Provides a way to freeze and unfreeze all tasks in a |
1221 | cgroup. |
1222 | |
1223 | This option affects the ORIGINAL cgroup interface. The cgroup2 memory |
1224 | controller includes important in-kernel memory consumers per default. |
1225 | |
1226 | If you're using cgroup2, say N. |
1227 | |
1228 | config CGROUP_HUGETLB |
1229 | bool "HugeTLB controller" |
1230 | depends on HUGETLB_PAGE |
1231 | select PAGE_COUNTER |
1232 | default n |
1233 | help |
1234 | Provides a cgroup controller for HugeTLB pages. |
1235 | When you enable this, you can put a per cgroup limit on HugeTLB usage. |
1236 | The limit is enforced during page fault. Since HugeTLB doesn't |
1237 | support page reclaim, enforcing the limit at page fault time implies |
1238 | that, the application will get SIGBUS signal if it tries to access |
1239 | HugeTLB pages beyond its limit. This requires the application to know |
1240 | beforehand how much HugeTLB pages it would require for its use. The |
1241 | control group is tracked in the third page lru pointer. This means |
1242 | that we cannot use the controller with huge page less than 3 pages. |
1243 | |
1244 | config CPUSETS |
1245 | bool "Cpuset controller" |
1246 | help |
1247 | This option will let you create and manage CPUSETs which |
1248 | allow dynamically partitioning a system into sets of CPUs and |
1249 | Memory Nodes and assigning tasks to run only within those sets. |
1250 | This is primarily useful on large SMP or NUMA systems. |
1251 | |
1252 | Say N if unsure. |
1253 | |
1254 | config PROC_PID_CPUSET |
1255 | bool "Include legacy /proc/<pid>/cpuset file" |
1256 | depends on CPUSETS |
1257 | default y |
1258 | |
1259 | config CGROUP_DEVICE |
1260 | bool "Device controller" |
1261 | help |
1262 | Provides a cgroup controller implementing whitelists for |
1263 | devices which a process in the cgroup can mknod or open. |
1264 | |
1265 | config CGROUP_CPUACCT |
1266 | bool "Simple CPU accounting controller" |
1267 | help |
1268 | Provides a simple controller for monitoring the |
1269 | total CPU consumed by the tasks in a cgroup. |
1270 | |
1271 | config CGROUP_PERF |
1272 | bool "Perf controller" |
1273 | depends on PERF_EVENTS |
1274 | help |
1275 | This option extends the perf per-cpu mode to restrict monitoring |
1276 | to threads which belong to the cgroup specified and run on the |
1277 | designated cpu. |
1278 | |
1279 | Say N if unsure. |
1280 | |
1281 | config CGROUP_BPF |
1282 | bool "Support for eBPF programs attached to cgroups" |
1283 | depends on BPF_SYSCALL |
1284 | select SOCK_CGROUP_DATA |
1285 | help |
1286 | Allow attaching eBPF programs to a cgroup using the bpf(2) |
1287 | syscall command BPF_PROG_ATTACH. |
1288 | |
1289 | In which context these programs are accessed depends on the type |
1290 | of attachment. For instance, programs that are attached using |
1291 | BPF_CGROUP_INET_INGRESS will be executed on the ingress path of |
1292 | inet sockets. |
1293 | |
1294 | config CGROUP_DEBUG |
1295 | bool "Example controller" |
1296 | default n |
1297 | help |
1298 | This option enables a simple controller that exports |
1299 | debugging information about the cgroups framework. |
1300 | |
1301 | Say N. |
1302 | |
1303 | config SOCK_CGROUP_DATA |
1304 | bool |
1305 | default n |
1306 | |
1307 | endif # CGROUPS |
1308 | |
1309 | config CHECKPOINT_RESTORE |
1310 | bool "Checkpoint/restore support" if EXPERT |
1311 | select PROC_CHILDREN |
1312 | default n |
1313 | help |
1314 | Enables additional kernel features in a sake of checkpoint/restore. |
1315 | In particular it adds auxiliary prctl codes to setup process text, |
1316 | data and heap segment sizes, and a few additional /proc filesystem |
1317 | entries. |
1318 | |
1319 | If unsure, say N here. |
1320 | |
1321 | menuconfig NAMESPACES |
1322 | bool "Namespaces support" if EXPERT |
1323 | depends on MULTIUSER |
1324 | default !EXPERT |
1325 | help |
1326 | Provides the way to make tasks work with different objects using |
1327 | the same id. For example same IPC id may refer to different objects |
1328 | or same user id or pid may refer to different tasks when used in |
1329 | different namespaces. |
1330 | |
1331 | if NAMESPACES |
1332 | |
1333 | config UTS_NS |
1334 | bool "UTS namespace" |
1335 | default y |
1336 | help |
1337 | In this namespace tasks see different info provided with the |
1338 | uname() system call |
1339 | |
1340 | config IPC_NS |
1341 | bool "IPC namespace" |
1342 | depends on (SYSVIPC || POSIX_MQUEUE) |
1343 | default y |
1344 | help |
1345 | In this namespace tasks work with IPC ids which correspond to |
1346 | different IPC objects in different namespaces. |
1347 | |
1348 | config USER_NS |
1349 | bool "User namespace" |
1350 | default n |
1351 | help |
1352 | This allows containers, i.e. vservers, to use user namespaces |
1353 | to provide different user info for different servers. |
1354 | |
1355 | When user namespaces are enabled in the kernel it is |
1356 | recommended that the MEMCG option also be enabled and that |
1357 | user-space use the memory control groups to limit the amount |
1358 | of memory a memory unprivileged users can use. |
1359 | |
1360 | If unsure, say N. |
1361 | |
1362 | config PID_NS |
1363 | bool "PID Namespaces" |
1364 | default y |
1365 | help |
1366 | Support process id namespaces. This allows having multiple |
1367 | processes with the same pid as long as they are in different |
1368 | pid namespaces. This is a building block of containers. |
1369 | |
1370 | config NET_NS |
1371 | bool "Network namespace" |
1372 | depends on NET |
1373 | default y |
1374 | help |
1375 | Allow user space to create what appear to be multiple instances |
1376 | of the network stack. |
1377 | |
1378 | endif # NAMESPACES |
1379 | |
1380 | config SCHED_AUTOGROUP |
1381 | bool "Automatic process group scheduling" |
1382 | select CGROUPS |
1383 | select CGROUP_SCHED |
1384 | select FAIR_GROUP_SCHED |
1385 | help |
1386 | This option optimizes the scheduler for common desktop workloads by |
1387 | automatically creating and populating task groups. This separation |
1388 | of workloads isolates aggressive CPU burners (like build jobs) from |
1389 | desktop applications. Task group autogeneration is currently based |
1390 | upon task session. |
1391 | |
1392 | config SCHED_TUNE |
1393 | bool "Boosting for CFS tasks (EXPERIMENTAL)" |
1394 | depends on SMP |
1395 | help |
1396 | This option enables the system-wide support for task boosting. |
1397 | When this support is enabled a new sysctl interface is exposed to |
1398 | userspace via: |
1399 | /proc/sys/kernel/sched_cfs_boost |
1400 | which allows to set a system-wide boost value in range [0..100]. |
1401 | |
1402 | The currently boosting strategy is implemented in such a way that: |
1403 | - a 0% boost value requires to operate in "standard" mode by |
1404 | scheduling all tasks at the minimum capacities required by their |
1405 | workload demand |
1406 | - a 100% boost value requires to push at maximum the task |
1407 | performances, "regardless" of the incurred energy consumption |
1408 | |
1409 | A boost value in between these two boundaries is used to bias the |
1410 | power/performance trade-off, the higher the boost value the more the |
1411 | scheduler is biased toward performance boosting instead of energy |
1412 | efficiency. |
1413 | |
1414 | Since this support exposes a single system-wide knob, the specified |
1415 | boost value is applied to all (CFS) tasks in the system. |
1416 | |
1417 | If unsure, say N. |
1418 | |
1419 | config DEFAULT_USE_ENERGY_AWARE |
1420 | bool "Default to enabling the Energy Aware Scheduler feature" |
1421 | default n |
1422 | help |
1423 | This option defaults the ENERGY_AWARE scheduling feature to true, |
1424 | as without SCHED_DEBUG set this feature can't be enabled or disabled |
1425 | via sysctl. |
1426 | |
1427 | Say N if unsure. |
1428 | |
1429 | config SYSFS_DEPRECATED |
1430 | bool "Enable deprecated sysfs features to support old userspace tools" |
1431 | depends on SYSFS |
1432 | default n |
1433 | help |
1434 | This option adds code that switches the layout of the "block" class |
1435 | devices, to not show up in /sys/class/block/, but only in |
1436 | /sys/block/. |
1437 | |
1438 | This switch is only active when the sysfs.deprecated=1 boot option is |
1439 | passed or the SYSFS_DEPRECATED_V2 option is set. |
1440 | |
1441 | This option allows new kernels to run on old distributions and tools, |
1442 | which might get confused by /sys/class/block/. Since 2007/2008 all |
1443 | major distributions and tools handle this just fine. |
1444 | |
1445 | Recent distributions and userspace tools after 2009/2010 depend on |
1446 | the existence of /sys/class/block/, and will not work with this |
1447 | option enabled. |
1448 | |
1449 | Only if you are using a new kernel on an old distribution, you might |
1450 | need to say Y here. |
1451 | |
1452 | config SYSFS_DEPRECATED_V2 |
1453 | bool "Enable deprecated sysfs features by default" |
1454 | default n |
1455 | depends on SYSFS |
1456 | depends on SYSFS_DEPRECATED |
1457 | help |
1458 | Enable deprecated sysfs by default. |
1459 | |
1460 | See the CONFIG_SYSFS_DEPRECATED option for more details about this |
1461 | option. |
1462 | |
1463 | Only if you are using a new kernel on an old distribution, you might |
1464 | need to say Y here. Even then, odds are you would not need it |
1465 | enabled, you can always pass the boot option if absolutely necessary. |
1466 | |
1467 | config RELAY |
1468 | bool "Kernel->user space relay support (formerly relayfs)" |
1469 | select IRQ_WORK |
1470 | help |
1471 | This option enables support for relay interface support in |
1472 | certain file systems (such as debugfs). |
1473 | It is designed to provide an efficient mechanism for tools and |
1474 | facilities to relay large amounts of data from kernel space to |
1475 | user space. |
1476 | |
1477 | If unsure, say N. |
1478 | |
1479 | config BLK_DEV_INITRD |
1480 | bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support" |
1481 | depends on BROKEN || !FRV |
1482 | help |
1483 | The initial RAM filesystem is a ramfs which is loaded by the |
1484 | boot loader (loadlin or lilo) and that is mounted as root |
1485 | before the normal boot procedure. It is typically used to |
1486 | load modules needed to mount the "real" root file system, |
1487 | etc. See <file:Documentation/initrd.txt> for details. |
1488 | |
1489 | If RAM disk support (BLK_DEV_RAM) is also included, this |
1490 | also enables initial RAM disk (initrd) support and adds |
1491 | 15 Kbytes (more on some other architectures) to the kernel size. |
1492 | |
1493 | If unsure say Y. |
1494 | |
1495 | if BLK_DEV_INITRD |
1496 | |
1497 | source "usr/Kconfig" |
1498 | |
1499 | endif |
1500 | |
1501 | choice |
1502 | prompt "Compiler optimization level" |
1503 | default CONFIG_CC_OPTIMIZE_FOR_PERFORMANCE |
1504 | |
1505 | config CC_OPTIMIZE_FOR_PERFORMANCE |
1506 | bool "Optimize for performance" |
1507 | help |
1508 | This is the default optimization level for the kernel, building |
1509 | with the "-O2" compiler flag for best performance and most |
1510 | helpful compile-time warnings. |
1511 | |
1512 | config CC_OPTIMIZE_FOR_SIZE |
1513 | bool "Optimize for size" |
1514 | help |
1515 | Enabling this option will pass "-Os" instead of "-O2" to |
1516 | your compiler resulting in a smaller kernel. |
1517 | |
1518 | If unsure, say N. |
1519 | |
1520 | endchoice |
1521 | |
1522 | config SYSCTL |
1523 | bool |
1524 | |
1525 | config ANON_INODES |
1526 | bool |
1527 | |
1528 | config HAVE_UID16 |
1529 | bool |
1530 | |
1531 | config SYSCTL_EXCEPTION_TRACE |
1532 | bool |
1533 | help |
1534 | Enable support for /proc/sys/debug/exception-trace. |
1535 | |
1536 | config SYSCTL_ARCH_UNALIGN_NO_WARN |
1537 | bool |
1538 | help |
1539 | Enable support for /proc/sys/kernel/ignore-unaligned-usertrap |
1540 | Allows arch to define/use @no_unaligned_warning to possibly warn |
1541 | about unaligned access emulation going on under the hood. |
1542 | |
1543 | config SYSCTL_ARCH_UNALIGN_ALLOW |
1544 | bool |
1545 | help |
1546 | Enable support for /proc/sys/kernel/unaligned-trap |
1547 | Allows arches to define/use @unaligned_enabled to runtime toggle |
1548 | the unaligned access emulation. |
1549 | see arch/parisc/kernel/unaligned.c for reference |
1550 | |
1551 | config HAVE_PCSPKR_PLATFORM |
1552 | bool |
1553 | |
1554 | # interpreter that classic socket filters depend on |
1555 | config BPF |
1556 | bool |
1557 | |
1558 | menuconfig EXPERT |
1559 | bool "Configure standard kernel features (expert users)" |
1560 | # Unhide debug options, to make the on-by-default options visible |
1561 | select DEBUG_KERNEL |
1562 | help |
1563 | This option allows certain base kernel options and settings |
1564 | to be disabled or tweaked. This is for specialized |
1565 | environments which can tolerate a "non-standard" kernel. |
1566 | Only use this if you really know what you are doing. |
1567 | |
1568 | config UID16 |
1569 | bool "Enable 16-bit UID system calls" if EXPERT |
1570 | depends on HAVE_UID16 && MULTIUSER |
1571 | default y |
1572 | help |
1573 | This enables the legacy 16-bit UID syscall wrappers. |
1574 | |
1575 | config MULTIUSER |
1576 | bool "Multiple users, groups and capabilities support" if EXPERT |
1577 | default y |
1578 | help |
1579 | This option enables support for non-root users, groups and |
1580 | capabilities. |
1581 | |
1582 | If you say N here, all processes will run with UID 0, GID 0, and all |
1583 | possible capabilities. Saying N here also compiles out support for |
1584 | system calls related to UIDs, GIDs, and capabilities, such as setuid, |
1585 | setgid, and capset. |
1586 | |
1587 | If unsure, say Y here. |
1588 | |
1589 | config SGETMASK_SYSCALL |
1590 | bool "sgetmask/ssetmask syscalls support" if EXPERT |
1591 | def_bool PARISC || MN10300 || BLACKFIN || M68K || PPC || MIPS || X86 || SPARC || CRIS || MICROBLAZE || SUPERH |
1592 | ---help--- |
1593 | sys_sgetmask and sys_ssetmask are obsolete system calls |
1594 | no longer supported in libc but still enabled by default in some |
1595 | architectures. |
1596 | |
1597 | If unsure, leave the default option here. |
1598 | |
1599 | config SYSFS_SYSCALL |
1600 | bool "Sysfs syscall support" if EXPERT |
1601 | default y |
1602 | ---help--- |
1603 | sys_sysfs is an obsolete system call no longer supported in libc. |
1604 | Note that disabling this option is more secure but might break |
1605 | compatibility with some systems. |
1606 | |
1607 | If unsure say Y here. |
1608 | |
1609 | config SYSCTL_SYSCALL |
1610 | bool "Sysctl syscall support" if EXPERT |
1611 | depends on PROC_SYSCTL |
1612 | default n |
1613 | select SYSCTL |
1614 | ---help--- |
1615 | sys_sysctl uses binary paths that have been found challenging |
1616 | to properly maintain and use. The interface in /proc/sys |
1617 | using paths with ascii names is now the primary path to this |
1618 | information. |
1619 | |
1620 | Almost nothing using the binary sysctl interface so if you are |
1621 | trying to save some space it is probably safe to disable this, |
1622 | making your kernel marginally smaller. |
1623 | |
1624 | If unsure say N here. |
1625 | |
1626 | config KALLSYMS |
1627 | bool "Load all symbols for debugging/ksymoops" if EXPERT |
1628 | default y |
1629 | help |
1630 | Say Y here to let the kernel print out symbolic crash information and |
1631 | symbolic stack backtraces. This increases the size of the kernel |
1632 | somewhat, as all symbols have to be loaded into the kernel image. |
1633 | |
1634 | config KALLSYMS_ALL |
1635 | bool "Include all symbols in kallsyms" |
1636 | depends on DEBUG_KERNEL && KALLSYMS |
1637 | help |
1638 | Normally kallsyms only contains the symbols of functions for nicer |
1639 | OOPS messages and backtraces (i.e., symbols from the text and inittext |
1640 | sections). This is sufficient for most cases. And only in very rare |
1641 | cases (e.g., when a debugger is used) all symbols are required (e.g., |
1642 | names of variables from the data sections, etc). |
1643 | |
1644 | This option makes sure that all symbols are loaded into the kernel |
1645 | image (i.e., symbols from all sections) in cost of increased kernel |
1646 | size (depending on the kernel configuration, it may be 300KiB or |
1647 | something like this). |
1648 | |
1649 | Say N unless you really need all symbols. |
1650 | |
1651 | config KALLSYMS_ABSOLUTE_PERCPU |
1652 | bool |
1653 | depends on KALLSYMS |
1654 | default X86_64 && SMP |
1655 | |
1656 | config KALLSYMS_BASE_RELATIVE |
1657 | bool |
1658 | depends on KALLSYMS |
1659 | default !IA64 && !(TILE && 64BIT) |
1660 | help |
1661 | Instead of emitting them as absolute values in the native word size, |
1662 | emit the symbol references in the kallsyms table as 32-bit entries, |
1663 | each containing a relative value in the range [base, base + U32_MAX] |
1664 | or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either |
1665 | an absolute value in the range [0, S32_MAX] or a relative value in the |
1666 | range [base, base + S32_MAX], where base is the lowest relative symbol |
1667 | address encountered in the image. |
1668 | |
1669 | On 64-bit builds, this reduces the size of the address table by 50%, |
1670 | but more importantly, it results in entries whose values are build |
1671 | time constants, and no relocation pass is required at runtime to fix |
1672 | up the entries based on the runtime load address of the kernel. |
1673 | |
1674 | config PRINTK |
1675 | default y |
1676 | bool "Enable support for printk" if EXPERT |
1677 | select IRQ_WORK |
1678 | help |
1679 | This option enables normal printk support. Removing it |
1680 | eliminates most of the message strings from the kernel image |
1681 | and makes the kernel more or less silent. As this makes it |
1682 | very difficult to diagnose system problems, saying N here is |
1683 | strongly discouraged. |
1684 | |
1685 | config PRINTK_NMI |
1686 | def_bool y |
1687 | depends on PRINTK |
1688 | depends on HAVE_NMI |
1689 | |
1690 | config BUG |
1691 | bool "BUG() support" if EXPERT |
1692 | default y |
1693 | help |
1694 | Disabling this option eliminates support for BUG and WARN, reducing |
1695 | the size of your kernel image and potentially quietly ignoring |
1696 | numerous fatal conditions. You should only consider disabling this |
1697 | option for embedded systems with no facilities for reporting errors. |
1698 | Just say Y. |
1699 | |
1700 | config ELF_CORE |
1701 | depends on COREDUMP |
1702 | default y |
1703 | bool "Enable ELF core dumps" if EXPERT |
1704 | help |
1705 | Enable support for generating core dumps. Disabling saves about 4k. |
1706 | |
1707 | |
1708 | config PCSPKR_PLATFORM |
1709 | bool "Enable PC-Speaker support" if EXPERT |
1710 | depends on HAVE_PCSPKR_PLATFORM |
1711 | select I8253_LOCK |
1712 | default y |
1713 | help |
1714 | This option allows to disable the internal PC-Speaker |
1715 | support, saving some memory. |
1716 | |
1717 | config BASE_FULL |
1718 | default y |
1719 | bool "Enable full-sized data structures for core" if EXPERT |
1720 | help |
1721 | Disabling this option reduces the size of miscellaneous core |
1722 | kernel data structures. This saves memory on small machines, |
1723 | but may reduce performance. |
1724 | |
1725 | config FUTEX |
1726 | bool "Enable futex support" if EXPERT |
1727 | default y |
1728 | select RT_MUTEXES |
1729 | help |
1730 | Disabling this option will cause the kernel to be built without |
1731 | support for "fast userspace mutexes". The resulting kernel may not |
1732 | run glibc-based applications correctly. |
1733 | |
1734 | config HAVE_FUTEX_CMPXCHG |
1735 | bool |
1736 | depends on FUTEX |
1737 | help |
1738 | Architectures should select this if futex_atomic_cmpxchg_inatomic() |
1739 | is implemented and always working. This removes a couple of runtime |
1740 | checks. |
1741 | |
1742 | config EPOLL |
1743 | bool "Enable eventpoll support" if EXPERT |
1744 | default y |
1745 | select ANON_INODES |
1746 | help |
1747 | Disabling this option will cause the kernel to be built without |
1748 | support for epoll family of system calls. |
1749 | |
1750 | config SIGNALFD |
1751 | bool "Enable signalfd() system call" if EXPERT |
1752 | select ANON_INODES |
1753 | default y |
1754 | help |
1755 | Enable the signalfd() system call that allows to receive signals |
1756 | on a file descriptor. |
1757 | |
1758 | If unsure, say Y. |
1759 | |
1760 | config TIMERFD |
1761 | bool "Enable timerfd() system call" if EXPERT |
1762 | select ANON_INODES |
1763 | default y |
1764 | help |
1765 | Enable the timerfd() system call that allows to receive timer |
1766 | events on a file descriptor. |
1767 | |
1768 | If unsure, say Y. |
1769 | |
1770 | config EVENTFD |
1771 | bool "Enable eventfd() system call" if EXPERT |
1772 | select ANON_INODES |
1773 | default y |
1774 | help |
1775 | Enable the eventfd() system call that allows to receive both |
1776 | kernel notification (ie. KAIO) or userspace notifications. |
1777 | |
1778 | If unsure, say Y. |
1779 | |
1780 | # syscall, maps, verifier |
1781 | config BPF_SYSCALL |
1782 | bool "Enable bpf() system call" |
1783 | select ANON_INODES |
1784 | select BPF |
1785 | default n |
1786 | help |
1787 | Enable the bpf() system call that allows to manipulate eBPF |
1788 | programs and maps via file descriptors. |
1789 | |
1790 | config BPF_JIT_ALWAYS_ON |
1791 | bool "Permanently enable BPF JIT and remove BPF interpreter" |
1792 | depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT |
1793 | help |
1794 | Enables BPF JIT and removes BPF interpreter to avoid |
1795 | speculative execution of BPF instructions by the interpreter |
1796 | |
1797 | config SHMEM |
1798 | bool "Use full shmem filesystem" if EXPERT |
1799 | default y |
1800 | depends on MMU |
1801 | help |
1802 | The shmem is an internal filesystem used to manage shared memory. |
1803 | It is backed by swap and manages resource limits. It is also exported |
1804 | to userspace as tmpfs if TMPFS is enabled. Disabling this |
1805 | option replaces shmem and tmpfs with the much simpler ramfs code, |
1806 | which may be appropriate on small systems without swap. |
1807 | |
1808 | config AIO |
1809 | bool "Enable AIO support" if EXPERT |
1810 | default y |
1811 | help |
1812 | This option enables POSIX asynchronous I/O which may by used |
1813 | by some high performance threaded applications. Disabling |
1814 | this option saves about 7k. |
1815 | |
1816 | config ADVISE_SYSCALLS |
1817 | bool "Enable madvise/fadvise syscalls" if EXPERT |
1818 | default y |
1819 | help |
1820 | This option enables the madvise and fadvise syscalls, used by |
1821 | applications to advise the kernel about their future memory or file |
1822 | usage, improving performance. If building an embedded system where no |
1823 | applications use these syscalls, you can disable this option to save |
1824 | space. |
1825 | |
1826 | config USERFAULTFD |
1827 | bool "Enable userfaultfd() system call" |
1828 | select ANON_INODES |
1829 | depends on MMU |
1830 | help |
1831 | Enable the userfaultfd() system call that allows to intercept and |
1832 | handle page faults in userland. |
1833 | |
1834 | config PCI_QUIRKS |
1835 | default y |
1836 | bool "Enable PCI quirk workarounds" if EXPERT |
1837 | depends on PCI |
1838 | help |
1839 | This enables workarounds for various PCI chipset |
1840 | bugs/quirks. Disable this only if your target machine is |
1841 | unaffected by PCI quirks. |
1842 | |
1843 | config MEMBARRIER |
1844 | bool "Enable membarrier() system call" if EXPERT |
1845 | default y |
1846 | help |
1847 | Enable the membarrier() system call that allows issuing memory |
1848 | barriers across all running threads, which can be used to distribute |
1849 | the cost of user-space memory barriers asymmetrically by transforming |
1850 | pairs of memory barriers into pairs consisting of membarrier() and a |
1851 | compiler barrier. |
1852 | |
1853 | If unsure, say Y. |
1854 | |
1855 | config EMBEDDED |
1856 | bool "Embedded system" |
1857 | option allnoconfig_y |
1858 | select EXPERT |
1859 | help |
1860 | This option should be enabled if compiling the kernel for |
1861 | an embedded system so certain expert options are available |
1862 | for configuration. |
1863 | |
1864 | config HAVE_PERF_EVENTS |
1865 | bool |
1866 | help |
1867 | See tools/perf/design.txt for details. |
1868 | |
1869 | config PERF_USE_VMALLOC |
1870 | bool |
1871 | help |
1872 | See tools/perf/design.txt for details |
1873 | |
1874 | menu "Kernel Performance Events And Counters" |
1875 | |
1876 | config PERF_EVENTS |
1877 | bool "Kernel performance events and counters" |
1878 | default y if PROFILING |
1879 | depends on HAVE_PERF_EVENTS |
1880 | select ANON_INODES |
1881 | select IRQ_WORK |
1882 | select SRCU |
1883 | help |
1884 | Enable kernel support for various performance events provided |
1885 | by software and hardware. |
1886 | |
1887 | Software events are supported either built-in or via the |
1888 | use of generic tracepoints. |
1889 | |
1890 | Most modern CPUs support performance events via performance |
1891 | counter registers. These registers count the number of certain |
1892 | types of hw events: such as instructions executed, cachemisses |
1893 | suffered, or branches mis-predicted - without slowing down the |
1894 | kernel or applications. These registers can also trigger interrupts |
1895 | when a threshold number of events have passed - and can thus be |
1896 | used to profile the code that runs on that CPU. |
1897 | |
1898 | The Linux Performance Event subsystem provides an abstraction of |
1899 | these software and hardware event capabilities, available via a |
1900 | system call and used by the "perf" utility in tools/perf/. It |
1901 | provides per task and per CPU counters, and it provides event |
1902 | capabilities on top of those. |
1903 | |
1904 | Say Y if unsure. |
1905 | |
1906 | config DEBUG_PERF_USE_VMALLOC |
1907 | default n |
1908 | bool "Debug: use vmalloc to back perf mmap() buffers" |
1909 | depends on PERF_EVENTS && DEBUG_KERNEL && !PPC |
1910 | select PERF_USE_VMALLOC |
1911 | help |
1912 | Use vmalloc memory to back perf mmap() buffers. |
1913 | |
1914 | Mostly useful for debugging the vmalloc code on platforms |
1915 | that don't require it. |
1916 | |
1917 | Say N if unsure. |
1918 | |
1919 | endmenu |
1920 | |
1921 | config VM_EVENT_COUNTERS |
1922 | default y |
1923 | bool "Enable VM event counters for /proc/vmstat" if EXPERT |
1924 | help |
1925 | VM event counters are needed for event counts to be shown. |
1926 | This option allows the disabling of the VM event counters |
1927 | on EXPERT systems. /proc/vmstat will only show page counts |
1928 | if VM event counters are disabled. |
1929 | |
1930 | config SLUB_DEBUG |
1931 | default y |
1932 | bool "Enable SLUB debugging support" if EXPERT |
1933 | depends on SLUB && SYSFS |
1934 | help |
1935 | SLUB has extensive debug support features. Disabling these can |
1936 | result in significant savings in code size. This also disables |
1937 | SLUB sysfs support. /sys/slab will not exist and there will be |
1938 | no support for cache validation etc. |
1939 | |
1940 | config SLUB_MEMCG_SYSFS_ON |
1941 | default n |
1942 | bool "Enable memcg SLUB sysfs support by default" if EXPERT |
1943 | depends on SLUB && SYSFS && MEMCG |
1944 | help |
1945 | SLUB creates a directory under /sys/kernel/slab for each |
1946 | allocation cache to host info and debug files. If memory |
1947 | cgroup is enabled, each cache can have per memory cgroup |
1948 | caches. SLUB can create the same sysfs directories for these |
1949 | caches under /sys/kernel/slab/CACHE/cgroup but it can lead |
1950 | to a very high number of debug files being created. This is |
1951 | controlled by slub_memcg_sysfs boot parameter and this |
1952 | config option determines the parameter's default value. |
1953 | |
1954 | config COMPAT_BRK |
1955 | bool "Disable heap randomization" |
1956 | default y |
1957 | help |
1958 | Randomizing heap placement makes heap exploits harder, but it |
1959 | also breaks ancient binaries (including anything libc5 based). |
1960 | This option changes the bootup default to heap randomization |
1961 | disabled, and can be overridden at runtime by setting |
1962 | /proc/sys/kernel/randomize_va_space to 2. |
1963 | |
1964 | On non-ancient distros (post-2000 ones) N is usually a safe choice. |
1965 | |
1966 | choice |
1967 | prompt "Choose SLAB allocator" |
1968 | default SLUB |
1969 | help |
1970 | This option allows to select a slab allocator. |
1971 | |
1972 | config SLAB |
1973 | bool "SLAB" |
1974 | select HAVE_HARDENED_USERCOPY_ALLOCATOR |
1975 | help |
1976 | The regular slab allocator that is established and known to work |
1977 | well in all environments. It organizes cache hot objects in |
1978 | per cpu and per node queues. |
1979 | |
1980 | config SLUB |
1981 | bool "SLUB (Unqueued Allocator)" |
1982 | select HAVE_HARDENED_USERCOPY_ALLOCATOR |
1983 | help |
1984 | SLUB is a slab allocator that minimizes cache line usage |
1985 | instead of managing queues of cached objects (SLAB approach). |
1986 | Per cpu caching is realized using slabs of objects instead |
1987 | of queues of objects. SLUB can use memory efficiently |
1988 | and has enhanced diagnostics. SLUB is the default choice for |
1989 | a slab allocator. |
1990 | |
1991 | config SLOB |
1992 | depends on EXPERT |
1993 | bool "SLOB (Simple Allocator)" |
1994 | help |
1995 | SLOB replaces the stock allocator with a drastically simpler |
1996 | allocator. SLOB is generally more space efficient but |
1997 | does not perform as well on large systems. |
1998 | |
1999 | endchoice |
2000 | |
2001 | config SLAB_FREELIST_RANDOM |
2002 | default n |
2003 | depends on SLAB || SLUB |
2004 | bool "SLAB freelist randomization" |
2005 | help |
2006 | Randomizes the freelist order used on creating new pages. This |
2007 | security feature reduces the predictability of the kernel slab |
2008 | allocator against heap overflows. |
2009 | |
2010 | config SLUB_CPU_PARTIAL |
2011 | default y |
2012 | depends on SLUB && SMP |
2013 | bool "SLUB per cpu partial cache" |
2014 | help |
2015 | Per cpu partial caches accellerate objects allocation and freeing |
2016 | that is local to a processor at the price of more indeterminism |
2017 | in the latency of the free. On overflow these caches will be cleared |
2018 | which requires the taking of locks that may cause latency spikes. |
2019 | Typically one would choose no for a realtime system. |
2020 | |
2021 | config MMAP_ALLOW_UNINITIALIZED |
2022 | bool "Allow mmapped anonymous memory to be uninitialized" |
2023 | depends on EXPERT && !MMU |
2024 | default n |
2025 | help |
2026 | Normally, and according to the Linux spec, anonymous memory obtained |
2027 | from mmap() has it's contents cleared before it is passed to |
2028 | userspace. Enabling this config option allows you to request that |
2029 | mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus |
2030 | providing a huge performance boost. If this option is not enabled, |
2031 | then the flag will be ignored. |
2032 | |
2033 | This is taken advantage of by uClibc's malloc(), and also by |
2034 | ELF-FDPIC binfmt's brk and stack allocator. |
2035 | |
2036 | Because of the obvious security issues, this option should only be |
2037 | enabled on embedded devices where you control what is run in |
2038 | userspace. Since that isn't generally a problem on no-MMU systems, |
2039 | it is normally safe to say Y here. |
2040 | |
2041 | See Documentation/nommu-mmap.txt for more information. |
2042 | |
2043 | config SYSTEM_DATA_VERIFICATION |
2044 | def_bool n |
2045 | select SYSTEM_TRUSTED_KEYRING |
2046 | select KEYS |
2047 | select CRYPTO |
2048 | select CRYPTO_RSA |
2049 | select ASYMMETRIC_KEY_TYPE |
2050 | select ASYMMETRIC_PUBLIC_KEY_SUBTYPE |
2051 | select ASN1 |
2052 | select OID_REGISTRY |
2053 | select X509_CERTIFICATE_PARSER |
2054 | select PKCS7_MESSAGE_PARSER |
2055 | help |
2056 | Provide PKCS#7 message verification using the contents of the system |
2057 | trusted keyring to provide public keys. This then can be used for |
2058 | module verification, kexec image verification and firmware blob |
2059 | verification. |
2060 | |
2061 | config PROFILING |
2062 | bool "Profiling support" |
2063 | help |
2064 | Say Y here to enable the extended profiling support mechanisms used |
2065 | by profilers such as OProfile. |
2066 | |
2067 | # |
2068 | # Place an empty function call at each tracepoint site. Can be |
2069 | # dynamically changed for a probe function. |
2070 | # |
2071 | config TRACEPOINTS |
2072 | bool |
2073 | |
2074 | source "arch/Kconfig" |
2075 | |
2076 | endmenu # General setup |
2077 | |
2078 | config HAVE_GENERIC_DMA_COHERENT |
2079 | bool |
2080 | default n |
2081 | |
2082 | config SLABINFO |
2083 | bool |
2084 | depends on PROC_FS |
2085 | depends on SLAB || SLUB_DEBUG |
2086 | default y |
2087 | |
2088 | config RT_MUTEXES |
2089 | bool |
2090 | |
2091 | config BASE_SMALL |
2092 | int |
2093 | default 0 if BASE_FULL |
2094 | default 1 if !BASE_FULL |
2095 | |
2096 | menuconfig MODULES |
2097 | bool "Enable loadable module support" |
2098 | option modules |
2099 | help |
2100 | Kernel modules are small pieces of compiled code which can |
2101 | be inserted in the running kernel, rather than being |
2102 | permanently built into the kernel. You use the "modprobe" |
2103 | tool to add (and sometimes remove) them. If you say Y here, |
2104 | many parts of the kernel can be built as modules (by |
2105 | answering M instead of Y where indicated): this is most |
2106 | useful for infrequently used options which are not required |
2107 | for booting. For more information, see the man pages for |
2108 | modprobe, lsmod, modinfo, insmod and rmmod. |
2109 | |
2110 | If you say Y here, you will need to run "make |
2111 | modules_install" to put the modules under /lib/modules/ |
2112 | where modprobe can find them (you may need to be root to do |
2113 | this). |
2114 | |
2115 | If unsure, say Y. |
2116 | |
2117 | if MODULES |
2118 | |
2119 | config MODULE_FORCE_LOAD |
2120 | bool "Forced module loading" |
2121 | default n |
2122 | help |
2123 | Allow loading of modules without version information (ie. modprobe |
2124 | --force). Forced module loading sets the 'F' (forced) taint flag and |
2125 | is usually a really bad idea. |
2126 | |
2127 | config MODULE_UNLOAD |
2128 | bool "Module unloading" |
2129 | help |
2130 | Without this option you will not be able to unload any |
2131 | modules (note that some modules may not be unloadable |
2132 | anyway), which makes your kernel smaller, faster |
2133 | and simpler. If unsure, say Y. |
2134 | |
2135 | config MODULE_FORCE_UNLOAD |
2136 | bool "Forced module unloading" |
2137 | depends on MODULE_UNLOAD |
2138 | help |
2139 | This option allows you to force a module to unload, even if the |
2140 | kernel believes it is unsafe: the kernel will remove the module |
2141 | without waiting for anyone to stop using it (using the -f option to |
2142 | rmmod). This is mainly for kernel developers and desperate users. |
2143 | If unsure, say N. |
2144 | |
2145 | config MODVERSIONS |
2146 | bool "Module versioning support" |
2147 | help |
2148 | Usually, you have to use modules compiled with your kernel. |
2149 | Saying Y here makes it sometimes possible to use modules |
2150 | compiled for different kernels, by adding enough information |
2151 | to the modules to (hopefully) spot any changes which would |
2152 | make them incompatible with the kernel you are running. If |
2153 | unsure, say N. |
2154 | |
2155 | config MODULE_SRCVERSION_ALL |
2156 | bool "Source checksum for all modules" |
2157 | help |
2158 | Modules which contain a MODULE_VERSION get an extra "srcversion" |
2159 | field inserted into their modinfo section, which contains a |
2160 | sum of the source files which made it. This helps maintainers |
2161 | see exactly which source was used to build a module (since |
2162 | others sometimes change the module source without updating |
2163 | the version). With this option, such a "srcversion" field |
2164 | will be created for all modules. If unsure, say N. |
2165 | |
2166 | config MODULE_SIG |
2167 | bool "Module signature verification" |
2168 | depends on MODULES |
2169 | select SYSTEM_DATA_VERIFICATION |
2170 | help |
2171 | Check modules for valid signatures upon load: the signature |
2172 | is simply appended to the module. For more information see |
2173 | Documentation/module-signing.txt. |
2174 | |
2175 | Note that this option adds the OpenSSL development packages as a |
2176 | kernel build dependency so that the signing tool can use its crypto |
2177 | library. |
2178 | |
2179 | !!!WARNING!!! If you enable this option, you MUST make sure that the |
2180 | module DOES NOT get stripped after being signed. This includes the |
2181 | debuginfo strip done by some packagers (such as rpmbuild) and |
2182 | inclusion into an initramfs that wants the module size reduced. |
2183 | |
2184 | config MODULE_SIG_FORCE |
2185 | bool "Require modules to be validly signed" |
2186 | depends on MODULE_SIG |
2187 | help |
2188 | Reject unsigned modules or signed modules for which we don't have a |
2189 | key. Without this, such modules will simply taint the kernel. |
2190 | |
2191 | config MODULE_SIG_ALL |
2192 | bool "Automatically sign all modules" |
2193 | default y |
2194 | depends on MODULE_SIG |
2195 | help |
2196 | Sign all modules during make modules_install. Without this option, |
2197 | modules must be signed manually, using the scripts/sign-file tool. |
2198 | |
2199 | comment "Do not forget to sign required modules with scripts/sign-file" |
2200 | depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL |
2201 | |
2202 | choice |
2203 | prompt "Which hash algorithm should modules be signed with?" |
2204 | depends on MODULE_SIG |
2205 | help |
2206 | This determines which sort of hashing algorithm will be used during |
2207 | signature generation. This algorithm _must_ be built into the kernel |
2208 | directly so that signature verification can take place. It is not |
2209 | possible to load a signed module containing the algorithm to check |
2210 | the signature on that module. |
2211 | |
2212 | config MODULE_SIG_SHA1 |
2213 | bool "Sign modules with SHA-1" |
2214 | select CRYPTO_SHA1 |
2215 | |
2216 | config MODULE_SIG_SHA224 |
2217 | bool "Sign modules with SHA-224" |
2218 | select CRYPTO_SHA256 |
2219 | |
2220 | config MODULE_SIG_SHA256 |
2221 | bool "Sign modules with SHA-256" |
2222 | select CRYPTO_SHA256 |
2223 | |
2224 | config MODULE_SIG_SHA384 |
2225 | bool "Sign modules with SHA-384" |
2226 | select CRYPTO_SHA512 |
2227 | |
2228 | config MODULE_SIG_SHA512 |
2229 | bool "Sign modules with SHA-512" |
2230 | select CRYPTO_SHA512 |
2231 | |
2232 | endchoice |
2233 | |
2234 | config MODULE_SIG_HASH |
2235 | string |
2236 | depends on MODULE_SIG |
2237 | default "sha1" if MODULE_SIG_SHA1 |
2238 | default "sha224" if MODULE_SIG_SHA224 |
2239 | default "sha256" if MODULE_SIG_SHA256 |
2240 | default "sha384" if MODULE_SIG_SHA384 |
2241 | default "sha512" if MODULE_SIG_SHA512 |
2242 | |
2243 | config MODULE_COMPRESS |
2244 | bool "Compress modules on installation" |
2245 | depends on MODULES |
2246 | help |
2247 | |
2248 | Compresses kernel modules when 'make modules_install' is run; gzip or |
2249 | xz depending on "Compression algorithm" below. |
2250 | |
2251 | module-init-tools MAY support gzip, and kmod MAY support gzip and xz. |
2252 | |
2253 | Out-of-tree kernel modules installed using Kbuild will also be |
2254 | compressed upon installation. |
2255 | |
2256 | Note: for modules inside an initrd or initramfs, it's more efficient |
2257 | to compress the whole initrd or initramfs instead. |
2258 | |
2259 | Note: This is fully compatible with signed modules. |
2260 | |
2261 | If in doubt, say N. |
2262 | |
2263 | choice |
2264 | prompt "Compression algorithm" |
2265 | depends on MODULE_COMPRESS |
2266 | default MODULE_COMPRESS_GZIP |
2267 | help |
2268 | This determines which sort of compression will be used during |
2269 | 'make modules_install'. |
2270 | |
2271 | GZIP (default) and XZ are supported. |
2272 | |
2273 | config MODULE_COMPRESS_GZIP |
2274 | bool "GZIP" |
2275 | |
2276 | config MODULE_COMPRESS_XZ |
2277 | bool "XZ" |
2278 | |
2279 | endchoice |
2280 | |
2281 | config TRIM_UNUSED_KSYMS |
2282 | bool "Trim unused exported kernel symbols" |
2283 | depends on MODULES && !UNUSED_SYMBOLS |
2284 | help |
2285 | The kernel and some modules make many symbols available for |
2286 | other modules to use via EXPORT_SYMBOL() and variants. Depending |
2287 | on the set of modules being selected in your kernel configuration, |
2288 | many of those exported symbols might never be used. |
2289 | |
2290 | This option allows for unused exported symbols to be dropped from |
2291 | the build. In turn, this provides the compiler more opportunities |
2292 | (especially when using LTO) for optimizing the code and reducing |
2293 | binary size. This might have some security advantages as well. |
2294 | |
2295 | If unsure, or if you need to build out-of-tree modules, say N. |
2296 | |
2297 | endif # MODULES |
2298 | |
2299 | config MODULES_TREE_LOOKUP |
2300 | def_bool y |
2301 | depends on PERF_EVENTS || TRACING || CFI_CLANG |
2302 | |
2303 | config INIT_ALL_POSSIBLE |
2304 | bool |
2305 | help |
2306 | Back when each arch used to define their own cpu_online_mask and |
2307 | cpu_possible_mask, some of them chose to initialize cpu_possible_mask |
2308 | with all 1s, and others with all 0s. When they were centralised, |
2309 | it was better to provide this option than to break all the archs |
2310 | and have several arch maintainers pursuing me down dark alleys. |
2311 | |
2312 | source "block/Kconfig" |
2313 | |
2314 | config PREEMPT_NOTIFIERS |
2315 | bool |
2316 | |
2317 | config PADATA |
2318 | depends on SMP |
2319 | bool |
2320 | |
2321 | config ASN1 |
2322 | tristate |
2323 | help |
2324 | Build a simple ASN.1 grammar compiler that produces a bytecode output |
2325 | that can be interpreted by the ASN.1 stream decoder and used to |
2326 | inform it as to what tags are to be expected in a stream and what |
2327 | functions to call on what tags. |
2328 | |
2329 | source "kernel/Kconfig.locks" |
2330 |