]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blob - kernel/sysctl.c
projects / mirror_ubuntu-artful-kernel.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
blame | history | raw | HEAD
nilfs2: use set_mask_bits() for operations on buffer state bitmap
[mirror_ubuntu-artful-kernel.git] / kernel / sysctl.c
1 /*
2 * sysctl.c: General linux system control interface
3 *
4 * Begun 24 March 1995, Stephen Tweedie
5 * Added /proc support, Dec 1995
6 * Added bdflush entry and intvec min/max checking, 2/23/96, Tom Dyas.
7 * Added hooks for /proc/sys/net (minor, minor patch), 96/4/1, Mike Shaver.
8 * Added kernel/java-{interpreter,appletviewer}, 96/5/10, Mike Shaver.
9 * Dynamic registration fixes, Stephen Tweedie.
10 * Added kswapd-interval, ctrl-alt-del, printk stuff, 1/8/97, Chris Horn.
11 * Made sysctl support optional via CONFIG_SYSCTL, 1/10/97, Chris
12 * Horn.
13 * Added proc_doulongvec_ms_jiffies_minmax, 09/08/99, Carlos H. Bauer.
14 * Added proc_doulongvec_minmax, 09/08/99, Carlos H. Bauer.
15 * Changed linked lists to use list.h instead of lists.h, 02/24/00, Bill
16 * Wendling.
17 * The list_for_each() macro wasn't appropriate for the sysctl loop.
18 * Removed it and replaced it with older style, 03/23/00, Bill Wendling
19 */
20
21 #include <linux/module.h>
22 #include <linux/aio.h>
23 #include <linux/mm.h>
24 #include <linux/swap.h>
25 #include <linux/slab.h>
26 #include <linux/sysctl.h>
27 #include <linux/bitmap.h>
28 #include <linux/signal.h>
29 #include <linux/printk.h>
30 #include <linux/proc_fs.h>
31 #include <linux/security.h>
32 #include <linux/ctype.h>
33 #include <linux/kmemcheck.h>
34 #include <linux/kmemleak.h>
35 #include <linux/fs.h>
36 #include <linux/init.h>
37 #include <linux/kernel.h>
38 #include <linux/kobject.h>
39 #include <linux/net.h>
40 #include <linux/sysrq.h>
41 #include <linux/highuid.h>
42 #include <linux/writeback.h>
43 #include <linux/ratelimit.h>
44 #include <linux/compaction.h>
45 #include <linux/hugetlb.h>
46 #include <linux/initrd.h>
47 #include <linux/key.h>
48 #include <linux/times.h>
49 #include <linux/limits.h>
50 #include <linux/dcache.h>
51 #include <linux/dnotify.h>
52 #include <linux/syscalls.h>
53 #include <linux/vmstat.h>
54 #include <linux/nfs_fs.h>
55 #include <linux/acpi.h>
56 #include <linux/reboot.h>
57 #include <linux/ftrace.h>
58 #include <linux/perf_event.h>
59 #include <linux/kprobes.h>
60 #include <linux/pipe_fs_i.h>
61 #include <linux/oom.h>
62 #include <linux/kmod.h>
63 #include <linux/capability.h>
64 #include <linux/binfmts.h>
65 #include <linux/sched/sysctl.h>
66 #include <linux/kexec.h>
67
68 #include <asm/uaccess.h>
69 #include <asm/processor.h>
70
71 #ifdef CONFIG_X86
72 #include <asm/nmi.h>
73 #include <asm/stacktrace.h>
74 #include <asm/io.h>
75 #endif
76 #ifdef CONFIG_SPARC
77 #include <asm/setup.h>
78 #endif
79 #ifdef CONFIG_BSD_PROCESS_ACCT
80 #include <linux/acct.h>
81 #endif
82 #ifdef CONFIG_RT_MUTEXES
83 #include <linux/rtmutex.h>
84 #endif
85 #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
86 #include <linux/lockdep.h>
87 #endif
88 #ifdef CONFIG_CHR_DEV_SG
89 #include <scsi/sg.h>
90 #endif
91
92 #ifdef CONFIG_LOCKUP_DETECTOR
93 #include <linux/nmi.h>
94 #endif
95
96
97 #if defined(CONFIG_SYSCTL)
98
99 /* External variables not in a header file. */
100 extern int max_threads;
101 extern int suid_dumpable;
102 #ifdef CONFIG_COREDUMP
103 extern int core_uses_pid;
104 extern char core_pattern[];
105 extern unsigned int core_pipe_limit;
106 #endif
107 extern int pid_max;
108 extern int pid_max_min, pid_max_max;
109 extern int percpu_pagelist_fraction;
110 extern int compat_log;
111 extern int latencytop_enabled;
112 extern int sysctl_nr_open_min, sysctl_nr_open_max;
113 #ifndef CONFIG_MMU
114 extern int sysctl_nr_trim_pages;
115 #endif
116
117 /* Constants used for minimum and maximum */
118 #ifdef CONFIG_LOCKUP_DETECTOR
119 static int sixty = 60;
120 #endif
121
122 static int __maybe_unused neg_one = -1;
123
124 static int zero;
125 static int __maybe_unused one = 1;
126 static int __maybe_unused two = 2;
127 static int __maybe_unused four = 4;
128 static unsigned long one_ul = 1;
129 static int one_hundred = 100;
130 #ifdef CONFIG_PRINTK
131 static int ten_thousand = 10000;
132 #endif
133
134 /* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
135 static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
136
137 /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
138 static int maxolduid = 65535;
139 static int minolduid;
140
141 static int ngroups_max = NGROUPS_MAX;
142 static const int cap_last_cap = CAP_LAST_CAP;
143
144 /*this is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs */
145 #ifdef CONFIG_DETECT_HUNG_TASK
146 static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
147 #endif
148
149 #ifdef CONFIG_INOTIFY_USER
150 #include <linux/inotify.h>
151 #endif
152 #ifdef CONFIG_SPARC
153 #endif
154
155 #ifdef __hppa__
156 extern int pwrsw_enabled;
157 #endif
158
159 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
160 extern int unaligned_enabled;
161 #endif
162
163 #ifdef CONFIG_IA64
164 extern int unaligned_dump_stack;
165 #endif
166
167 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
168 extern int no_unaligned_warning;
169 #endif
170
171 #ifdef CONFIG_PROC_SYSCTL
172
173 #define SYSCTL_WRITES_LEGACY -1
174 #define SYSCTL_WRITES_WARN 0
175 #define SYSCTL_WRITES_STRICT 1
176
177 static int sysctl_writes_strict = SYSCTL_WRITES_WARN;
178
179 static int proc_do_cad_pid(struct ctl_table *table, int write,
180 void __user *buffer, size_t *lenp, loff_t *ppos);
181 static int proc_taint(struct ctl_table *table, int write,
182 void __user *buffer, size_t *lenp, loff_t *ppos);
183 #endif
184
185 #ifdef CONFIG_PRINTK
186 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
187 void __user *buffer, size_t *lenp, loff_t *ppos);
188 #endif
189
190 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
191 void __user *buffer, size_t *lenp, loff_t *ppos);
192 #ifdef CONFIG_COREDUMP
193 static int proc_dostring_coredump(struct ctl_table *table, int write,
194 void __user *buffer, size_t *lenp, loff_t *ppos);
195 #endif
196
197 #ifdef CONFIG_MAGIC_SYSRQ
198 /* Note: sysrq code uses it's own private copy */
199 static int __sysrq_enabled = CONFIG_MAGIC_SYSRQ_DEFAULT_ENABLE;
200
201 static int sysrq_sysctl_handler(struct ctl_table *table, int write,
202 void __user *buffer, size_t *lenp,
203 loff_t *ppos)
204 {
205 int error;
206
207 error = proc_dointvec(table, write, buffer, lenp, ppos);
208 if (error)
209 return error;
210
211 if (write)
212 sysrq_toggle_support(__sysrq_enabled);
213
214 return 0;
215 }
216
217 #endif
218
219 static struct ctl_table kern_table[];
220 static struct ctl_table vm_table[];
221 static struct ctl_table fs_table[];
222 static struct ctl_table debug_table[];
223 static struct ctl_table dev_table[];
224 extern struct ctl_table random_table[];
225 #ifdef CONFIG_EPOLL
226 extern struct ctl_table epoll_table[];
227 #endif
228
229 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
230 int sysctl_legacy_va_layout;
231 #endif
232
233 /* The default sysctl tables: */
234
235 static struct ctl_table sysctl_base_table[] = {
236 {
237 .procname = "kernel",
238 .mode = 0555,
239 .child = kern_table,
240 },
241 {
242 .procname = "vm",
243 .mode = 0555,
244 .child = vm_table,
245 },
246 {
247 .procname = "fs",
248 .mode = 0555,
249 .child = fs_table,
250 },
251 {
252 .procname = "debug",
253 .mode = 0555,
254 .child = debug_table,
255 },
256 {
257 .procname = "dev",
258 .mode = 0555,
259 .child = dev_table,
260 },
261 { }
262 };
263
264 #ifdef CONFIG_SCHED_DEBUG
265 static int min_sched_granularity_ns = 100000; /* 100 usecs */
266 static int max_sched_granularity_ns = NSEC_PER_SEC; /* 1 second */
267 static int min_wakeup_granularity_ns; /* 0 usecs */
268 static int max_wakeup_granularity_ns = NSEC_PER_SEC; /* 1 second */
269 #ifdef CONFIG_SMP
270 static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
271 static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
272 #endif /* CONFIG_SMP */
273 #endif /* CONFIG_SCHED_DEBUG */
274
275 #ifdef CONFIG_COMPACTION
276 static int min_extfrag_threshold;
277 static int max_extfrag_threshold = 1000;
278 #endif
279
280 static struct ctl_table kern_table[] = {
281 {
282 .procname = "sched_child_runs_first",
283 .data = &sysctl_sched_child_runs_first,
284 .maxlen = sizeof(unsigned int),
285 .mode = 0644,
286 .proc_handler = proc_dointvec,
287 },
288 #ifdef CONFIG_SCHED_DEBUG
289 {
290 .procname = "sched_min_granularity_ns",
291 .data = &sysctl_sched_min_granularity,
292 .maxlen = sizeof(unsigned int),
293 .mode = 0644,
294 .proc_handler = sched_proc_update_handler,
295 .extra1 = &min_sched_granularity_ns,
296 .extra2 = &max_sched_granularity_ns,
297 },
298 {
299 .procname = "sched_latency_ns",
300 .data = &sysctl_sched_latency,
301 .maxlen = sizeof(unsigned int),
302 .mode = 0644,
303 .proc_handler = sched_proc_update_handler,
304 .extra1 = &min_sched_granularity_ns,
305 .extra2 = &max_sched_granularity_ns,
306 },
307 {
308 .procname = "sched_wakeup_granularity_ns",
309 .data = &sysctl_sched_wakeup_granularity,
310 .maxlen = sizeof(unsigned int),
311 .mode = 0644,
312 .proc_handler = sched_proc_update_handler,
313 .extra1 = &min_wakeup_granularity_ns,
314 .extra2 = &max_wakeup_granularity_ns,
315 },
316 #ifdef CONFIG_SMP
317 {
318 .procname = "sched_tunable_scaling",
319 .data = &sysctl_sched_tunable_scaling,
320 .maxlen = sizeof(enum sched_tunable_scaling),
321 .mode = 0644,
322 .proc_handler = sched_proc_update_handler,
323 .extra1 = &min_sched_tunable_scaling,
324 .extra2 = &max_sched_tunable_scaling,
325 },
326 {
327 .procname = "sched_migration_cost_ns",
328 .data = &sysctl_sched_migration_cost,
329 .maxlen = sizeof(unsigned int),
330 .mode = 0644,
331 .proc_handler = proc_dointvec,
332 },
333 {
334 .procname = "sched_nr_migrate",
335 .data = &sysctl_sched_nr_migrate,
336 .maxlen = sizeof(unsigned int),
337 .mode = 0644,
338 .proc_handler = proc_dointvec,
339 },
340 {
341 .procname = "sched_time_avg_ms",
342 .data = &sysctl_sched_time_avg,
343 .maxlen = sizeof(unsigned int),
344 .mode = 0644,
345 .proc_handler = proc_dointvec,
346 },
347 {
348 .procname = "sched_shares_window_ns",
349 .data = &sysctl_sched_shares_window,
350 .maxlen = sizeof(unsigned int),
351 .mode = 0644,
352 .proc_handler = proc_dointvec,
353 },
354 {
355 .procname = "timer_migration",
356 .data = &sysctl_timer_migration,
357 .maxlen = sizeof(unsigned int),
358 .mode = 0644,
359 .proc_handler = proc_dointvec_minmax,
360 .extra1 = &zero,
361 .extra2 = &one,
362 },
363 #endif /* CONFIG_SMP */
364 #ifdef CONFIG_NUMA_BALANCING
365 {
366 .procname = "numa_balancing_scan_delay_ms",
367 .data = &sysctl_numa_balancing_scan_delay,
368 .maxlen = sizeof(unsigned int),
369 .mode = 0644,
370 .proc_handler = proc_dointvec,
371 },
372 {
373 .procname = "numa_balancing_scan_period_min_ms",
374 .data = &sysctl_numa_balancing_scan_period_min,
375 .maxlen = sizeof(unsigned int),
376 .mode = 0644,
377 .proc_handler = proc_dointvec,
378 },
379 {
380 .procname = "numa_balancing_scan_period_max_ms",
381 .data = &sysctl_numa_balancing_scan_period_max,
382 .maxlen = sizeof(unsigned int),
383 .mode = 0644,
384 .proc_handler = proc_dointvec,
385 },
386 {
387 .procname = "numa_balancing_scan_size_mb",
388 .data = &sysctl_numa_balancing_scan_size,
389 .maxlen = sizeof(unsigned int),
390 .mode = 0644,
391 .proc_handler = proc_dointvec_minmax,
392 .extra1 = &one,
393 },
394 {
395 .procname = "numa_balancing",
396 .data = NULL, /* filled in by handler */
397 .maxlen = sizeof(unsigned int),
398 .mode = 0644,
399 .proc_handler = sysctl_numa_balancing,
400 .extra1 = &zero,
401 .extra2 = &one,
402 },
403 #endif /* CONFIG_NUMA_BALANCING */
404 #endif /* CONFIG_SCHED_DEBUG */
405 {
406 .procname = "sched_rt_period_us",
407 .data = &sysctl_sched_rt_period,
408 .maxlen = sizeof(unsigned int),
409 .mode = 0644,
410 .proc_handler = sched_rt_handler,
411 },
412 {
413 .procname = "sched_rt_runtime_us",
414 .data = &sysctl_sched_rt_runtime,
415 .maxlen = sizeof(int),
416 .mode = 0644,
417 .proc_handler = sched_rt_handler,
418 },
419 {
420 .procname = "sched_rr_timeslice_ms",
421 .data = &sched_rr_timeslice,
422 .maxlen = sizeof(int),
423 .mode = 0644,
424 .proc_handler = sched_rr_handler,
425 },
426 #ifdef CONFIG_SCHED_AUTOGROUP
427 {
428 .procname = "sched_autogroup_enabled",
429 .data = &sysctl_sched_autogroup_enabled,
430 .maxlen = sizeof(unsigned int),
431 .mode = 0644,
432 .proc_handler = proc_dointvec_minmax,
433 .extra1 = &zero,
434 .extra2 = &one,
435 },
436 #endif
437 #ifdef CONFIG_CFS_BANDWIDTH
438 {
439 .procname = "sched_cfs_bandwidth_slice_us",
440 .data = &sysctl_sched_cfs_bandwidth_slice,
441 .maxlen = sizeof(unsigned int),
442 .mode = 0644,
443 .proc_handler = proc_dointvec_minmax,
444 .extra1 = &one,
445 },
446 #endif
447 #ifdef CONFIG_PROVE_LOCKING
448 {
449 .procname = "prove_locking",
450 .data = &prove_locking,
451 .maxlen = sizeof(int),
452 .mode = 0644,
453 .proc_handler = proc_dointvec,
454 },
455 #endif
456 #ifdef CONFIG_LOCK_STAT
457 {
458 .procname = "lock_stat",
459 .data = &lock_stat,
460 .maxlen = sizeof(int),
461 .mode = 0644,
462 .proc_handler = proc_dointvec,
463 },
464 #endif
465 {
466 .procname = "panic",
467 .data = &panic_timeout,
468 .maxlen = sizeof(int),
469 .mode = 0644,
470 .proc_handler = proc_dointvec,
471 },
472 #ifdef CONFIG_COREDUMP
473 {
474 .procname = "core_uses_pid",
475 .data = &core_uses_pid,
476 .maxlen = sizeof(int),
477 .mode = 0644,
478 .proc_handler = proc_dointvec,
479 },
480 {
481 .procname = "core_pattern",
482 .data = core_pattern,
483 .maxlen = CORENAME_MAX_SIZE,
484 .mode = 0644,
485 .proc_handler = proc_dostring_coredump,
486 },
487 {
488 .procname = "core_pipe_limit",
489 .data = &core_pipe_limit,
490 .maxlen = sizeof(unsigned int),
491 .mode = 0644,
492 .proc_handler = proc_dointvec,
493 },
494 #endif
495 #ifdef CONFIG_PROC_SYSCTL
496 {
497 .procname = "tainted",
498 .maxlen = sizeof(long),
499 .mode = 0644,
500 .proc_handler = proc_taint,
501 },
502 {
503 .procname = "sysctl_writes_strict",
504 .data = &sysctl_writes_strict,
505 .maxlen = sizeof(int),
506 .mode = 0644,
507 .proc_handler = proc_dointvec_minmax,
508 .extra1 = &neg_one,
509 .extra2 = &one,
510 },
511 #endif
512 #ifdef CONFIG_LATENCYTOP
513 {
514 .procname = "latencytop",
515 .data = &latencytop_enabled,
516 .maxlen = sizeof(int),
517 .mode = 0644,
518 .proc_handler = proc_dointvec,
519 },
520 #endif
521 #ifdef CONFIG_BLK_DEV_INITRD
522 {
523 .procname = "real-root-dev",
524 .data = &real_root_dev,
525 .maxlen = sizeof(int),
526 .mode = 0644,
527 .proc_handler = proc_dointvec,
528 },
529 #endif
530 {
531 .procname = "print-fatal-signals",
532 .data = &print_fatal_signals,
533 .maxlen = sizeof(int),
534 .mode = 0644,
535 .proc_handler = proc_dointvec,
536 },
537 #ifdef CONFIG_SPARC
538 {
539 .procname = "reboot-cmd",
540 .data = reboot_command,
541 .maxlen = 256,
542 .mode = 0644,
543 .proc_handler = proc_dostring,
544 },
545 {
546 .procname = "stop-a",
547 .data = &stop_a_enabled,
548 .maxlen = sizeof (int),
549 .mode = 0644,
550 .proc_handler = proc_dointvec,
551 },
552 {
553 .procname = "scons-poweroff",
554 .data = &scons_pwroff,
555 .maxlen = sizeof (int),
556 .mode = 0644,
557 .proc_handler = proc_dointvec,
558 },
559 #endif
560 #ifdef CONFIG_SPARC64
561 {
562 .procname = "tsb-ratio",
563 .data = &sysctl_tsb_ratio,
564 .maxlen = sizeof (int),
565 .mode = 0644,
566 .proc_handler = proc_dointvec,
567 },
568 #endif
569 #ifdef __hppa__
570 {
571 .procname = "soft-power",
572 .data = &pwrsw_enabled,
573 .maxlen = sizeof (int),
574 .mode = 0644,
575 .proc_handler = proc_dointvec,
576 },
577 #endif
578 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
579 {
580 .procname = "unaligned-trap",
581 .data = &unaligned_enabled,
582 .maxlen = sizeof (int),
583 .mode = 0644,
584 .proc_handler = proc_dointvec,
585 },
586 #endif
587 {
588 .procname = "ctrl-alt-del",
589 .data = &C_A_D,
590 .maxlen = sizeof(int),
591 .mode = 0644,
592 .proc_handler = proc_dointvec,
593 },
594 #ifdef CONFIG_FUNCTION_TRACER
595 {
596 .procname = "ftrace_enabled",
597 .data = &ftrace_enabled,
598 .maxlen = sizeof(int),
599 .mode = 0644,
600 .proc_handler = ftrace_enable_sysctl,
601 },
602 #endif
603 #ifdef CONFIG_STACK_TRACER
604 {
605 .procname = "stack_tracer_enabled",
606 .data = &stack_tracer_enabled,
607 .maxlen = sizeof(int),
608 .mode = 0644,
609 .proc_handler = stack_trace_sysctl,
610 },
611 #endif
612 #ifdef CONFIG_TRACING
613 {
614 .procname = "ftrace_dump_on_oops",
615 .data = &ftrace_dump_on_oops,
616 .maxlen = sizeof(int),
617 .mode = 0644,
618 .proc_handler = proc_dointvec,
619 },
620 {
621 .procname = "traceoff_on_warning",
622 .data = &__disable_trace_on_warning,
623 .maxlen = sizeof(__disable_trace_on_warning),
624 .mode = 0644,
625 .proc_handler = proc_dointvec,
626 },
627 {
628 .procname = "tracepoint_printk",
629 .data = &tracepoint_printk,
630 .maxlen = sizeof(tracepoint_printk),
631 .mode = 0644,
632 .proc_handler = proc_dointvec,
633 },
634 #endif
635 #ifdef CONFIG_KEXEC
636 {
637 .procname = "kexec_load_disabled",
638 .data = &kexec_load_disabled,
639 .maxlen = sizeof(int),
640 .mode = 0644,
641 /* only handle a transition from default "0" to "1" */
642 .proc_handler = proc_dointvec_minmax,
643 .extra1 = &one,
644 .extra2 = &one,
645 },
646 #endif
647 #ifdef CONFIG_MODULES
648 {
649 .procname = "modprobe",
650 .data = &modprobe_path,
651 .maxlen = KMOD_PATH_LEN,
652 .mode = 0644,
653 .proc_handler = proc_dostring,
654 },
655 {
656 .procname = "modules_disabled",
657 .data = &modules_disabled,
658 .maxlen = sizeof(int),
659 .mode = 0644,
660 /* only handle a transition from default "0" to "1" */
661 .proc_handler = proc_dointvec_minmax,
662 .extra1 = &one,
663 .extra2 = &one,
664 },
665 #endif
666 #ifdef CONFIG_UEVENT_HELPER
667 {
668 .procname = "hotplug",
669 .data = &uevent_helper,
670 .maxlen = UEVENT_HELPER_PATH_LEN,
671 .mode = 0644,
672 .proc_handler = proc_dostring,
673 },
674 #endif
675 #ifdef CONFIG_CHR_DEV_SG
676 {
677 .procname = "sg-big-buff",
678 .data = &sg_big_buff,
679 .maxlen = sizeof (int),
680 .mode = 0444,
681 .proc_handler = proc_dointvec,
682 },
683 #endif
684 #ifdef CONFIG_BSD_PROCESS_ACCT
685 {
686 .procname = "acct",
687 .data = &acct_parm,
688 .maxlen = 3*sizeof(int),
689 .mode = 0644,
690 .proc_handler = proc_dointvec,
691 },
692 #endif
693 #ifdef CONFIG_MAGIC_SYSRQ
694 {
695 .procname = "sysrq",
696 .data = &__sysrq_enabled,
697 .maxlen = sizeof (int),
698 .mode = 0644,
699 .proc_handler = sysrq_sysctl_handler,
700 },
701 #endif
702 #ifdef CONFIG_PROC_SYSCTL
703 {
704 .procname = "cad_pid",
705 .data = NULL,
706 .maxlen = sizeof (int),
707 .mode = 0600,
708 .proc_handler = proc_do_cad_pid,
709 },
710 #endif
711 {
712 .procname = "threads-max",
713 .data = &max_threads,
714 .maxlen = sizeof(int),
715 .mode = 0644,
716 .proc_handler = proc_dointvec,
717 },
718 {
719 .procname = "random",
720 .mode = 0555,
721 .child = random_table,
722 },
723 {
724 .procname = "usermodehelper",
725 .mode = 0555,
726 .child = usermodehelper_table,
727 },
728 {
729 .procname = "overflowuid",
730 .data = &overflowuid,
731 .maxlen = sizeof(int),
732 .mode = 0644,
733 .proc_handler = proc_dointvec_minmax,
734 .extra1 = &minolduid,
735 .extra2 = &maxolduid,
736 },
737 {
738 .procname = "overflowgid",
739 .data = &overflowgid,
740 .maxlen = sizeof(int),
741 .mode = 0644,
742 .proc_handler = proc_dointvec_minmax,
743 .extra1 = &minolduid,
744 .extra2 = &maxolduid,
745 },
746 #ifdef CONFIG_S390
747 #ifdef CONFIG_MATHEMU
748 {
749 .procname = "ieee_emulation_warnings",
750 .data = &sysctl_ieee_emulation_warnings,
751 .maxlen = sizeof(int),
752 .mode = 0644,
753 .proc_handler = proc_dointvec,
754 },
755 #endif
756 {
757 .procname = "userprocess_debug",
758 .data = &show_unhandled_signals,
759 .maxlen = sizeof(int),
760 .mode = 0644,
761 .proc_handler = proc_dointvec,
762 },
763 #endif
764 {
765 .procname = "pid_max",
766 .data = &pid_max,
767 .maxlen = sizeof (int),
768 .mode = 0644,
769 .proc_handler = proc_dointvec_minmax,
770 .extra1 = &pid_max_min,
771 .extra2 = &pid_max_max,
772 },
773 {
774 .procname = "panic_on_oops",
775 .data = &panic_on_oops,
776 .maxlen = sizeof(int),
777 .mode = 0644,
778 .proc_handler = proc_dointvec,
779 },
780 #if defined CONFIG_PRINTK
781 {
782 .procname = "printk",
783 .data = &console_loglevel,
784 .maxlen = 4*sizeof(int),
785 .mode = 0644,
786 .proc_handler = proc_dointvec,
787 },
788 {
789 .procname = "printk_ratelimit",
790 .data = &printk_ratelimit_state.interval,
791 .maxlen = sizeof(int),
792 .mode = 0644,
793 .proc_handler = proc_dointvec_jiffies,
794 },
795 {
796 .procname = "printk_ratelimit_burst",
797 .data = &printk_ratelimit_state.burst,
798 .maxlen = sizeof(int),
799 .mode = 0644,
800 .proc_handler = proc_dointvec,
801 },
802 {
803 .procname = "printk_delay",
804 .data = &printk_delay_msec,
805 .maxlen = sizeof(int),
806 .mode = 0644,
807 .proc_handler = proc_dointvec_minmax,
808 .extra1 = &zero,
809 .extra2 = &ten_thousand,
810 },
811 {
812 .procname = "dmesg_restrict",
813 .data = &dmesg_restrict,
814 .maxlen = sizeof(int),
815 .mode = 0644,
816 .proc_handler = proc_dointvec_minmax_sysadmin,
817 .extra1 = &zero,
818 .extra2 = &one,
819 },
820 {
821 .procname = "kptr_restrict",
822 .data = &kptr_restrict,
823 .maxlen = sizeof(int),
824 .mode = 0644,
825 .proc_handler = proc_dointvec_minmax_sysadmin,
826 .extra1 = &zero,
827 .extra2 = &two,
828 },
829 #endif
830 {
831 .procname = "ngroups_max",
832 .data = &ngroups_max,
833 .maxlen = sizeof (int),
834 .mode = 0444,
835 .proc_handler = proc_dointvec,
836 },
837 {
838 .procname = "cap_last_cap",
839 .data = (void *)&cap_last_cap,
840 .maxlen = sizeof(int),
841 .mode = 0444,
842 .proc_handler = proc_dointvec,
843 },
844 #if defined(CONFIG_LOCKUP_DETECTOR)
845 {
846 .procname = "watchdog",
847 .data = &watchdog_user_enabled,
848 .maxlen = sizeof (int),
849 .mode = 0644,
850 .proc_handler = proc_watchdog,
851 .extra1 = &zero,
852 .extra2 = &one,
853 },
854 {
855 .procname = "watchdog_thresh",
856 .data = &watchdog_thresh,
857 .maxlen = sizeof(int),
858 .mode = 0644,
859 .proc_handler = proc_watchdog_thresh,
860 .extra1 = &zero,
861 .extra2 = &sixty,
862 },
863 {
864 .procname = "nmi_watchdog",
865 .data = &nmi_watchdog_enabled,
866 .maxlen = sizeof (int),
867 .mode = 0644,
868 .proc_handler = proc_nmi_watchdog,
869 .extra1 = &zero,
870 #if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR)
871 .extra2 = &one,
872 #else
873 .extra2 = &zero,
874 #endif
875 },
876 {
877 .procname = "soft_watchdog",
878 .data = &soft_watchdog_enabled,
879 .maxlen = sizeof (int),
880 .mode = 0644,
881 .proc_handler = proc_soft_watchdog,
882 .extra1 = &zero,
883 .extra2 = &one,
884 },
885 {
886 .procname = "softlockup_panic",
887 .data = &softlockup_panic,
888 .maxlen = sizeof(int),
889 .mode = 0644,
890 .proc_handler = proc_dointvec_minmax,
891 .extra1 = &zero,
892 .extra2 = &one,
893 },
894 #ifdef CONFIG_SMP
895 {
896 .procname = "softlockup_all_cpu_backtrace",
897 .data = &sysctl_softlockup_all_cpu_backtrace,
898 .maxlen = sizeof(int),
899 .mode = 0644,
900 .proc_handler = proc_dointvec_minmax,
901 .extra1 = &zero,
902 .extra2 = &one,
903 },
904 #endif /* CONFIG_SMP */
905 #endif
906 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
907 {
908 .procname = "unknown_nmi_panic",
909 .data = &unknown_nmi_panic,
910 .maxlen = sizeof (int),
911 .mode = 0644,
912 .proc_handler = proc_dointvec,
913 },
914 #endif
915 #if defined(CONFIG_X86)
916 {
917 .procname = "panic_on_unrecovered_nmi",
918 .data = &panic_on_unrecovered_nmi,
919 .maxlen = sizeof(int),
920 .mode = 0644,
921 .proc_handler = proc_dointvec,
922 },
923 {
924 .procname = "panic_on_io_nmi",
925 .data = &panic_on_io_nmi,
926 .maxlen = sizeof(int),
927 .mode = 0644,
928 .proc_handler = proc_dointvec,
929 },
930 #ifdef CONFIG_DEBUG_STACKOVERFLOW
931 {
932 .procname = "panic_on_stackoverflow",
933 .data = &sysctl_panic_on_stackoverflow,
934 .maxlen = sizeof(int),
935 .mode = 0644,
936 .proc_handler = proc_dointvec,
937 },
938 #endif
939 {
940 .procname = "bootloader_type",
941 .data = &bootloader_type,
942 .maxlen = sizeof (int),
943 .mode = 0444,
944 .proc_handler = proc_dointvec,
945 },
946 {
947 .procname = "bootloader_version",
948 .data = &bootloader_version,
949 .maxlen = sizeof (int),
950 .mode = 0444,
951 .proc_handler = proc_dointvec,
952 },
953 {
954 .procname = "kstack_depth_to_print",
955 .data = &kstack_depth_to_print,
956 .maxlen = sizeof(int),
957 .mode = 0644,
958 .proc_handler = proc_dointvec,
959 },
960 {
961 .procname = "io_delay_type",
962 .data = &io_delay_type,
963 .maxlen = sizeof(int),
964 .mode = 0644,
965 .proc_handler = proc_dointvec,
966 },
967 #endif
968 #if defined(CONFIG_MMU)
969 {
970 .procname = "randomize_va_space",
971 .data = &randomize_va_space,
972 .maxlen = sizeof(int),
973 .mode = 0644,
974 .proc_handler = proc_dointvec,
975 },
976 #endif
977 #if defined(CONFIG_S390) && defined(CONFIG_SMP)
978 {
979 .procname = "spin_retry",
980 .data = &spin_retry,
981 .maxlen = sizeof (int),
982 .mode = 0644,
983 .proc_handler = proc_dointvec,
984 },
985 #endif
986 #if defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
987 {
988 .procname = "acpi_video_flags",
989 .data = &acpi_realmode_flags,
990 .maxlen = sizeof (unsigned long),
991 .mode = 0644,
992 .proc_handler = proc_doulongvec_minmax,
993 },
994 #endif
995 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
996 {
997 .procname = "ignore-unaligned-usertrap",
998 .data = &no_unaligned_warning,
999 .maxlen = sizeof (int),
1000 .mode = 0644,
1001 .proc_handler = proc_dointvec,
1002 },
1003 #endif
1004 #ifdef CONFIG_IA64
1005 {
1006 .procname = "unaligned-dump-stack",
1007 .data = &unaligned_dump_stack,
1008 .maxlen = sizeof (int),
1009 .mode = 0644,
1010 .proc_handler = proc_dointvec,
1011 },
1012 #endif
1013 #ifdef CONFIG_DETECT_HUNG_TASK
1014 {
1015 .procname = "hung_task_panic",
1016 .data = &sysctl_hung_task_panic,
1017 .maxlen = sizeof(int),
1018 .mode = 0644,
1019 .proc_handler = proc_dointvec_minmax,
1020 .extra1 = &zero,
1021 .extra2 = &one,
1022 },
1023 {
1024 .procname = "hung_task_check_count",
1025 .data = &sysctl_hung_task_check_count,
1026 .maxlen = sizeof(int),
1027 .mode = 0644,
1028 .proc_handler = proc_dointvec_minmax,
1029 .extra1 = &zero,
1030 },
1031 {
1032 .procname = "hung_task_timeout_secs",
1033 .data = &sysctl_hung_task_timeout_secs,
1034 .maxlen = sizeof(unsigned long),
1035 .mode = 0644,
1036 .proc_handler = proc_dohung_task_timeout_secs,
1037 .extra2 = &hung_task_timeout_max,
1038 },
1039 {
1040 .procname = "hung_task_warnings",
1041 .data = &sysctl_hung_task_warnings,
1042 .maxlen = sizeof(int),
1043 .mode = 0644,
1044 .proc_handler = proc_dointvec_minmax,
1045 .extra1 = &neg_one,
1046 },
1047 #endif
1048 #ifdef CONFIG_COMPAT
1049 {
1050 .procname = "compat-log",
1051 .data = &compat_log,
1052 .maxlen = sizeof (int),
1053 .mode = 0644,
1054 .proc_handler = proc_dointvec,
1055 },
1056 #endif
1057 #ifdef CONFIG_RT_MUTEXES
1058 {
1059 .procname = "max_lock_depth",
1060 .data = &max_lock_depth,
1061 .maxlen = sizeof(int),
1062 .mode = 0644,
1063 .proc_handler = proc_dointvec,
1064 },
1065 #endif
1066 {
1067 .procname = "poweroff_cmd",
1068 .data = &poweroff_cmd,
1069 .maxlen = POWEROFF_CMD_PATH_LEN,
1070 .mode = 0644,
1071 .proc_handler = proc_dostring,
1072 },
1073 #ifdef CONFIG_KEYS
1074 {
1075 .procname = "keys",
1076 .mode = 0555,
1077 .child = key_sysctls,
1078 },
1079 #endif
1080 #ifdef CONFIG_PERF_EVENTS
1081 /*
1082 * User-space scripts rely on the existence of this file
1083 * as a feature check for perf_events being enabled.
1084 *
1085 * So it's an ABI, do not remove!
1086 */
1087 {
1088 .procname = "perf_event_paranoid",
1089 .data = &sysctl_perf_event_paranoid,
1090 .maxlen = sizeof(sysctl_perf_event_paranoid),
1091 .mode = 0644,
1092 .proc_handler = proc_dointvec,
1093 },
1094 {
1095 .procname = "perf_event_mlock_kb",
1096 .data = &sysctl_perf_event_mlock,
1097 .maxlen = sizeof(sysctl_perf_event_mlock),
1098 .mode = 0644,
1099 .proc_handler = proc_dointvec,
1100 },
1101 {
1102 .procname = "perf_event_max_sample_rate",
1103 .data = &sysctl_perf_event_sample_rate,
1104 .maxlen = sizeof(sysctl_perf_event_sample_rate),
1105 .mode = 0644,
1106 .proc_handler = perf_proc_update_handler,
1107 .extra1 = &one,
1108 },
1109 {
1110 .procname = "perf_cpu_time_max_percent",
1111 .data = &sysctl_perf_cpu_time_max_percent,
1112 .maxlen = sizeof(sysctl_perf_cpu_time_max_percent),
1113 .mode = 0644,
1114 .proc_handler = perf_cpu_time_max_percent_handler,
1115 .extra1 = &zero,
1116 .extra2 = &one_hundred,
1117 },
1118 #endif
1119 #ifdef CONFIG_KMEMCHECK
1120 {
1121 .procname = "kmemcheck",
1122 .data = &kmemcheck_enabled,
1123 .maxlen = sizeof(int),
1124 .mode = 0644,
1125 .proc_handler = proc_dointvec,
1126 },
1127 #endif
1128 {
1129 .procname = "panic_on_warn",
1130 .data = &panic_on_warn,
1131 .maxlen = sizeof(int),
1132 .mode = 0644,
1133 .proc_handler = proc_dointvec_minmax,
1134 .extra1 = &zero,
1135 .extra2 = &one,
1136 },
1137 { }
1138 };
1139
1140 static struct ctl_table vm_table[] = {
1141 {
1142 .procname = "overcommit_memory",
1143 .data = &sysctl_overcommit_memory,
1144 .maxlen = sizeof(sysctl_overcommit_memory),
1145 .mode = 0644,
1146 .proc_handler = proc_dointvec_minmax,
1147 .extra1 = &zero,
1148 .extra2 = &two,
1149 },
1150 {
1151 .procname = "panic_on_oom",
1152 .data = &sysctl_panic_on_oom,
1153 .maxlen = sizeof(sysctl_panic_on_oom),
1154 .mode = 0644,
1155 .proc_handler = proc_dointvec_minmax,
1156 .extra1 = &zero,
1157 .extra2 = &two,
1158 },
1159 {
1160 .procname = "oom_kill_allocating_task",
1161 .data = &sysctl_oom_kill_allocating_task,
1162 .maxlen = sizeof(sysctl_oom_kill_allocating_task),
1163 .mode = 0644,
1164 .proc_handler = proc_dointvec,
1165 },
1166 {
1167 .procname = "oom_dump_tasks",
1168 .data = &sysctl_oom_dump_tasks,
1169 .maxlen = sizeof(sysctl_oom_dump_tasks),
1170 .mode = 0644,
1171 .proc_handler = proc_dointvec,
1172 },
1173 {
1174 .procname = "overcommit_ratio",
1175 .data = &sysctl_overcommit_ratio,
1176 .maxlen = sizeof(sysctl_overcommit_ratio),
1177 .mode = 0644,
1178 .proc_handler = overcommit_ratio_handler,
1179 },
1180 {
1181 .procname = "overcommit_kbytes",
1182 .data = &sysctl_overcommit_kbytes,
1183 .maxlen = sizeof(sysctl_overcommit_kbytes),
1184 .mode = 0644,
1185 .proc_handler = overcommit_kbytes_handler,
1186 },
1187 {
1188 .procname = "page-cluster",
1189 .data = &page_cluster,
1190 .maxlen = sizeof(int),
1191 .mode = 0644,
1192 .proc_handler = proc_dointvec_minmax,
1193 .extra1 = &zero,
1194 },
1195 {
1196 .procname = "dirty_background_ratio",
1197 .data = &dirty_background_ratio,
1198 .maxlen = sizeof(dirty_background_ratio),
1199 .mode = 0644,
1200 .proc_handler = dirty_background_ratio_handler,
1201 .extra1 = &zero,
1202 .extra2 = &one_hundred,
1203 },
1204 {
1205 .procname = "dirty_background_bytes",
1206 .data = &dirty_background_bytes,
1207 .maxlen = sizeof(dirty_background_bytes),
1208 .mode = 0644,
1209 .proc_handler = dirty_background_bytes_handler,
1210 .extra1 = &one_ul,
1211 },
1212 {
1213 .procname = "dirty_ratio",
1214 .data = &vm_dirty_ratio,
1215 .maxlen = sizeof(vm_dirty_ratio),
1216 .mode = 0644,
1217 .proc_handler = dirty_ratio_handler,
1218 .extra1 = &zero,
1219 .extra2 = &one_hundred,
1220 },
1221 {
1222 .procname = "dirty_bytes",
1223 .data = &vm_dirty_bytes,
1224 .maxlen = sizeof(vm_dirty_bytes),
1225 .mode = 0644,
1226 .proc_handler = dirty_bytes_handler,
1227 .extra1 = &dirty_bytes_min,
1228 },
1229 {
1230 .procname = "dirty_writeback_centisecs",
1231 .data = &dirty_writeback_interval,
1232 .maxlen = sizeof(dirty_writeback_interval),
1233 .mode = 0644,
1234 .proc_handler = dirty_writeback_centisecs_handler,
1235 },
1236 {
1237 .procname = "dirty_expire_centisecs",
1238 .data = &dirty_expire_interval,
1239 .maxlen = sizeof(dirty_expire_interval),
1240 .mode = 0644,
1241 .proc_handler = proc_dointvec_minmax,
1242 .extra1 = &zero,
1243 },
1244 {
1245 .procname = "dirtytime_expire_seconds",
1246 .data = &dirtytime_expire_interval,
1247 .maxlen = sizeof(dirty_expire_interval),
1248 .mode = 0644,
1249 .proc_handler = dirtytime_interval_handler,
1250 .extra1 = &zero,
1251 },
1252 {
1253 .procname = "nr_pdflush_threads",
1254 .mode = 0444 /* read-only */,
1255 .proc_handler = pdflush_proc_obsolete,
1256 },
1257 {
1258 .procname = "swappiness",
1259 .data = &vm_swappiness,
1260 .maxlen = sizeof(vm_swappiness),
1261 .mode = 0644,
1262 .proc_handler = proc_dointvec_minmax,
1263 .extra1 = &zero,
1264 .extra2 = &one_hundred,
1265 },
1266 #ifdef CONFIG_HUGETLB_PAGE
1267 {
1268 .procname = "nr_hugepages",
1269 .data = NULL,
1270 .maxlen = sizeof(unsigned long),
1271 .mode = 0644,
1272 .proc_handler = hugetlb_sysctl_handler,
1273 },
1274 #ifdef CONFIG_NUMA
1275 {
1276 .procname = "nr_hugepages_mempolicy",
1277 .data = NULL,
1278 .maxlen = sizeof(unsigned long),
1279 .mode = 0644,
1280 .proc_handler = &hugetlb_mempolicy_sysctl_handler,
1281 },
1282 #endif
1283 {
1284 .procname = "hugetlb_shm_group",
1285 .data = &sysctl_hugetlb_shm_group,
1286 .maxlen = sizeof(gid_t),
1287 .mode = 0644,
1288 .proc_handler = proc_dointvec,
1289 },
1290 {
1291 .procname = "hugepages_treat_as_movable",
1292 .data = &hugepages_treat_as_movable,
1293 .maxlen = sizeof(int),
1294 .mode = 0644,
1295 .proc_handler = proc_dointvec,
1296 },
1297 {
1298 .procname = "nr_overcommit_hugepages",
1299 .data = NULL,
1300 .maxlen = sizeof(unsigned long),
1301 .mode = 0644,
1302 .proc_handler = hugetlb_overcommit_handler,
1303 },
1304 #endif
1305 {
1306 .procname = "lowmem_reserve_ratio",
1307 .data = &sysctl_lowmem_reserve_ratio,
1308 .maxlen = sizeof(sysctl_lowmem_reserve_ratio),
1309 .mode = 0644,
1310 .proc_handler = lowmem_reserve_ratio_sysctl_handler,
1311 },
1312 {
1313 .procname = "drop_caches",
1314 .data = &sysctl_drop_caches,
1315 .maxlen = sizeof(int),
1316 .mode = 0644,
1317 .proc_handler = drop_caches_sysctl_handler,
1318 .extra1 = &one,
1319 .extra2 = &four,
1320 },
1321 #ifdef CONFIG_COMPACTION
1322 {
1323 .procname = "compact_memory",
1324 .data = &sysctl_compact_memory,
1325 .maxlen = sizeof(int),
1326 .mode = 0200,
1327 .proc_handler = sysctl_compaction_handler,
1328 },
1329 {
1330 .procname = "extfrag_threshold",
1331 .data = &sysctl_extfrag_threshold,
1332 .maxlen = sizeof(int),
1333 .mode = 0644,
1334 .proc_handler = sysctl_extfrag_handler,
1335 .extra1 = &min_extfrag_threshold,
1336 .extra2 = &max_extfrag_threshold,
1337 },
1338 {
1339 .procname = "compact_unevictable_allowed",
1340 .data = &sysctl_compact_unevictable_allowed,
1341 .maxlen = sizeof(int),
1342 .mode = 0644,
1343 .proc_handler = proc_dointvec,
1344 .extra1 = &zero,
1345 .extra2 = &one,
1346 },
1347
1348 #endif /* CONFIG_COMPACTION */
1349 {
1350 .procname = "min_free_kbytes",
1351 .data = &min_free_kbytes,
1352 .maxlen = sizeof(min_free_kbytes),
1353 .mode = 0644,
1354 .proc_handler = min_free_kbytes_sysctl_handler,
1355 .extra1 = &zero,
1356 },
1357 {
1358 .procname = "percpu_pagelist_fraction",
1359 .data = &percpu_pagelist_fraction,
1360 .maxlen = sizeof(percpu_pagelist_fraction),
1361 .mode = 0644,
1362 .proc_handler = percpu_pagelist_fraction_sysctl_handler,
1363 .extra1 = &zero,
1364 },
1365 #ifdef CONFIG_MMU
1366 {
1367 .procname = "max_map_count",
1368 .data = &sysctl_max_map_count,
1369 .maxlen = sizeof(sysctl_max_map_count),
1370 .mode = 0644,
1371 .proc_handler = proc_dointvec_minmax,
1372 .extra1 = &zero,
1373 },
1374 #else
1375 {
1376 .procname = "nr_trim_pages",
1377 .data = &sysctl_nr_trim_pages,
1378 .maxlen = sizeof(sysctl_nr_trim_pages),
1379 .mode = 0644,
1380 .proc_handler = proc_dointvec_minmax,
1381 .extra1 = &zero,
1382 },
1383 #endif
1384 {
1385 .procname = "laptop_mode",
1386 .data = &laptop_mode,
1387 .maxlen = sizeof(laptop_mode),
1388 .mode = 0644,
1389 .proc_handler = proc_dointvec_jiffies,
1390 },
1391 {
1392 .procname = "block_dump",
1393 .data = &block_dump,
1394 .maxlen = sizeof(block_dump),
1395 .mode = 0644,
1396 .proc_handler = proc_dointvec,
1397 .extra1 = &zero,
1398 },
1399 {
1400 .procname = "vfs_cache_pressure",
1401 .data = &sysctl_vfs_cache_pressure,
1402 .maxlen = sizeof(sysctl_vfs_cache_pressure),
1403 .mode = 0644,
1404 .proc_handler = proc_dointvec,
1405 .extra1 = &zero,
1406 },
1407 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1408 {
1409 .procname = "legacy_va_layout",
1410 .data = &sysctl_legacy_va_layout,
1411 .maxlen = sizeof(sysctl_legacy_va_layout),
1412 .mode = 0644,
1413 .proc_handler = proc_dointvec,
1414 .extra1 = &zero,
1415 },
1416 #endif
1417 #ifdef CONFIG_NUMA
1418 {
1419 .procname = "zone_reclaim_mode",
1420 .data = &zone_reclaim_mode,
1421 .maxlen = sizeof(zone_reclaim_mode),
1422 .mode = 0644,
1423 .proc_handler = proc_dointvec,
1424 .extra1 = &zero,
1425 },
1426 {
1427 .procname = "min_unmapped_ratio",
1428 .data = &sysctl_min_unmapped_ratio,
1429 .maxlen = sizeof(sysctl_min_unmapped_ratio),
1430 .mode = 0644,
1431 .proc_handler = sysctl_min_unmapped_ratio_sysctl_handler,
1432 .extra1 = &zero,
1433 .extra2 = &one_hundred,
1434 },
1435 {
1436 .procname = "min_slab_ratio",
1437 .data = &sysctl_min_slab_ratio,
1438 .maxlen = sizeof(sysctl_min_slab_ratio),
1439 .mode = 0644,
1440 .proc_handler = sysctl_min_slab_ratio_sysctl_handler,
1441 .extra1 = &zero,
1442 .extra2 = &one_hundred,
1443 },
1444 #endif
1445 #ifdef CONFIG_SMP
1446 {
1447 .procname = "stat_interval",
1448 .data = &sysctl_stat_interval,
1449 .maxlen = sizeof(sysctl_stat_interval),
1450 .mode = 0644,
1451 .proc_handler = proc_dointvec_jiffies,
1452 },
1453 #endif
1454 #ifdef CONFIG_MMU
1455 {
1456 .procname = "mmap_min_addr",
1457 .data = &dac_mmap_min_addr,
1458 .maxlen = sizeof(unsigned long),
1459 .mode = 0644,
1460 .proc_handler = mmap_min_addr_handler,
1461 },
1462 #endif
1463 #ifdef CONFIG_NUMA
1464 {
1465 .procname = "numa_zonelist_order",
1466 .data = &numa_zonelist_order,
1467 .maxlen = NUMA_ZONELIST_ORDER_LEN,
1468 .mode = 0644,
1469 .proc_handler = numa_zonelist_order_handler,
1470 },
1471 #endif
1472 #if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
1473 (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
1474 {
1475 .procname = "vdso_enabled",
1476 #ifdef CONFIG_X86_32
1477 .data = &vdso32_enabled,
1478 .maxlen = sizeof(vdso32_enabled),
1479 #else
1480 .data = &vdso_enabled,
1481 .maxlen = sizeof(vdso_enabled),
1482 #endif
1483 .mode = 0644,
1484 .proc_handler = proc_dointvec,
1485 .extra1 = &zero,
1486 },
1487 #endif
1488 #ifdef CONFIG_HIGHMEM
1489 {
1490 .procname = "highmem_is_dirtyable",
1491 .data = &vm_highmem_is_dirtyable,
1492 .maxlen = sizeof(vm_highmem_is_dirtyable),
1493 .mode = 0644,
1494 .proc_handler = proc_dointvec_minmax,
1495 .extra1 = &zero,
1496 .extra2 = &one,
1497 },
1498 #endif
1499 #ifdef CONFIG_MEMORY_FAILURE
1500 {
1501 .procname = "memory_failure_early_kill",
1502 .data = &sysctl_memory_failure_early_kill,
1503 .maxlen = sizeof(sysctl_memory_failure_early_kill),
1504 .mode = 0644,
1505 .proc_handler = proc_dointvec_minmax,
1506 .extra1 = &zero,
1507 .extra2 = &one,
1508 },
1509 {
1510 .procname = "memory_failure_recovery",
1511 .data = &sysctl_memory_failure_recovery,
1512 .maxlen = sizeof(sysctl_memory_failure_recovery),
1513 .mode = 0644,
1514 .proc_handler = proc_dointvec_minmax,
1515 .extra1 = &zero,
1516 .extra2 = &one,
1517 },
1518 #endif
1519 {
1520 .procname = "user_reserve_kbytes",
1521 .data = &sysctl_user_reserve_kbytes,
1522 .maxlen = sizeof(sysctl_user_reserve_kbytes),
1523 .mode = 0644,
1524 .proc_handler = proc_doulongvec_minmax,
1525 },
1526 {
1527 .procname = "admin_reserve_kbytes",
1528 .data = &sysctl_admin_reserve_kbytes,
1529 .maxlen = sizeof(sysctl_admin_reserve_kbytes),
1530 .mode = 0644,
1531 .proc_handler = proc_doulongvec_minmax,
1532 },
1533 { }
1534 };
1535
1536 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1537 static struct ctl_table binfmt_misc_table[] = {
1538 { }
1539 };
1540 #endif
1541
1542 static struct ctl_table fs_table[] = {
1543 {
1544 .procname = "inode-nr",
1545 .data = &inodes_stat,
1546 .maxlen = 2*sizeof(long),
1547 .mode = 0444,
1548 .proc_handler = proc_nr_inodes,
1549 },
1550 {
1551 .procname = "inode-state",
1552 .data = &inodes_stat,
1553 .maxlen = 7*sizeof(long),
1554 .mode = 0444,
1555 .proc_handler = proc_nr_inodes,
1556 },
1557 {
1558 .procname = "file-nr",
1559 .data = &files_stat,
1560 .maxlen = sizeof(files_stat),
1561 .mode = 0444,
1562 .proc_handler = proc_nr_files,
1563 },
1564 {
1565 .procname = "file-max",
1566 .data = &files_stat.max_files,
1567 .maxlen = sizeof(files_stat.max_files),
1568 .mode = 0644,
1569 .proc_handler = proc_doulongvec_minmax,
1570 },
1571 {
1572 .procname = "nr_open",
1573 .data = &sysctl_nr_open,
1574 .maxlen = sizeof(int),
1575 .mode = 0644,
1576 .proc_handler = proc_dointvec_minmax,
1577 .extra1 = &sysctl_nr_open_min,
1578 .extra2 = &sysctl_nr_open_max,
1579 },
1580 {
1581 .procname = "dentry-state",
1582 .data = &dentry_stat,
1583 .maxlen = 6*sizeof(long),
1584 .mode = 0444,
1585 .proc_handler = proc_nr_dentry,
1586 },
1587 {
1588 .procname = "overflowuid",
1589 .data = &fs_overflowuid,
1590 .maxlen = sizeof(int),
1591 .mode = 0644,
1592 .proc_handler = proc_dointvec_minmax,
1593 .extra1 = &minolduid,
1594 .extra2 = &maxolduid,
1595 },
1596 {
1597 .procname = "overflowgid",
1598 .data = &fs_overflowgid,
1599 .maxlen = sizeof(int),
1600 .mode = 0644,
1601 .proc_handler = proc_dointvec_minmax,
1602 .extra1 = &minolduid,
1603 .extra2 = &maxolduid,
1604 },
1605 #ifdef CONFIG_FILE_LOCKING
1606 {
1607 .procname = "leases-enable",
1608 .data = &leases_enable,
1609 .maxlen = sizeof(int),
1610 .mode = 0644,
1611 .proc_handler = proc_dointvec,
1612 },
1613 #endif
1614 #ifdef CONFIG_DNOTIFY
1615 {
1616 .procname = "dir-notify-enable",
1617 .data = &dir_notify_enable,
1618 .maxlen = sizeof(int),
1619 .mode = 0644,
1620 .proc_handler = proc_dointvec,
1621 },
1622 #endif
1623 #ifdef CONFIG_MMU
1624 #ifdef CONFIG_FILE_LOCKING
1625 {
1626 .procname = "lease-break-time",
1627 .data = &lease_break_time,
1628 .maxlen = sizeof(int),
1629 .mode = 0644,
1630 .proc_handler = proc_dointvec,
1631 },
1632 #endif
1633 #ifdef CONFIG_AIO
1634 {
1635 .procname = "aio-nr",
1636 .data = &aio_nr,
1637 .maxlen = sizeof(aio_nr),
1638 .mode = 0444,
1639 .proc_handler = proc_doulongvec_minmax,
1640 },
1641 {
1642 .procname = "aio-max-nr",
1643 .data = &aio_max_nr,
1644 .maxlen = sizeof(aio_max_nr),
1645 .mode = 0644,
1646 .proc_handler = proc_doulongvec_minmax,
1647 },
1648 #endif /* CONFIG_AIO */
1649 #ifdef CONFIG_INOTIFY_USER
1650 {
1651 .procname = "inotify",
1652 .mode = 0555,
1653 .child = inotify_table,
1654 },
1655 #endif
1656 #ifdef CONFIG_EPOLL
1657 {
1658 .procname = "epoll",
1659 .mode = 0555,
1660 .child = epoll_table,
1661 },
1662 #endif
1663 #endif
1664 {
1665 .procname = "protected_symlinks",
1666 .data = &sysctl_protected_symlinks,
1667 .maxlen = sizeof(int),
1668 .mode = 0600,
1669 .proc_handler = proc_dointvec_minmax,
1670 .extra1 = &zero,
1671 .extra2 = &one,
1672 },
1673 {
1674 .procname = "protected_hardlinks",
1675 .data = &sysctl_protected_hardlinks,
1676 .maxlen = sizeof(int),
1677 .mode = 0600,
1678 .proc_handler = proc_dointvec_minmax,
1679 .extra1 = &zero,
1680 .extra2 = &one,
1681 },
1682 {
1683 .procname = "suid_dumpable",
1684 .data = &suid_dumpable,
1685 .maxlen = sizeof(int),
1686 .mode = 0644,
1687 .proc_handler = proc_dointvec_minmax_coredump,
1688 .extra1 = &zero,
1689 .extra2 = &two,
1690 },
1691 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1692 {
1693 .procname = "binfmt_misc",
1694 .mode = 0555,
1695 .child = binfmt_misc_table,
1696 },
1697 #endif
1698 {
1699 .procname = "pipe-max-size",
1700 .data = &pipe_max_size,
1701 .maxlen = sizeof(int),
1702 .mode = 0644,
1703 .proc_handler = &pipe_proc_fn,
1704 .extra1 = &pipe_min_size,
1705 },
1706 { }
1707 };
1708
1709 static struct ctl_table debug_table[] = {
1710 #ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
1711 {
1712 .procname = "exception-trace",
1713 .data = &show_unhandled_signals,
1714 .maxlen = sizeof(int),
1715 .mode = 0644,
1716 .proc_handler = proc_dointvec
1717 },
1718 #endif
1719 #if defined(CONFIG_OPTPROBES)
1720 {
1721 .procname = "kprobes-optimization",
1722 .data = &sysctl_kprobes_optimization,
1723 .maxlen = sizeof(int),
1724 .mode = 0644,
1725 .proc_handler = proc_kprobes_optimization_handler,
1726 .extra1 = &zero,
1727 .extra2 = &one,
1728 },
1729 #endif
1730 { }
1731 };
1732
1733 static struct ctl_table dev_table[] = {
1734 { }
1735 };
1736
1737 int __init sysctl_init(void)
1738 {
1739 struct ctl_table_header *hdr;
1740
1741 hdr = register_sysctl_table(sysctl_base_table);
1742 kmemleak_not_leak(hdr);
1743 return 0;
1744 }
1745
1746 #endif /* CONFIG_SYSCTL */
1747
1748 /*
1749 * /proc/sys support
1750 */
1751
1752 #ifdef CONFIG_PROC_SYSCTL
1753
1754 static int _proc_do_string(char *data, int maxlen, int write,
1755 char __user *buffer,
1756 size_t *lenp, loff_t *ppos)
1757 {
1758 size_t len;
1759 char __user *p;
1760 char c;
1761
1762 if (!data || !maxlen || !*lenp) {
1763 *lenp = 0;
1764 return 0;
1765 }
1766
1767 if (write) {
1768 if (sysctl_writes_strict == SYSCTL_WRITES_STRICT) {
1769 /* Only continue writes not past the end of buffer. */
1770 len = strlen(data);
1771 if (len > maxlen - 1)
1772 len = maxlen - 1;
1773
1774 if (*ppos > len)
1775 return 0;
1776 len = *ppos;
1777 } else {
1778 /* Start writing from beginning of buffer. */
1779 len = 0;
1780 }
1781
1782 *ppos += *lenp;
1783 p = buffer;
1784 while ((p - buffer) < *lenp && len < maxlen - 1) {
1785 if (get_user(c, p++))
1786 return -EFAULT;
1787 if (c == 0 || c == '\n')
1788 break;
1789 data[len++] = c;
1790 }
1791 data[len] = 0;
1792 } else {
1793 len = strlen(data);
1794 if (len > maxlen)
1795 len = maxlen;
1796
1797 if (*ppos > len) {
1798 *lenp = 0;
1799 return 0;
1800 }
1801
1802 data += *ppos;
1803 len -= *ppos;
1804
1805 if (len > *lenp)
1806 len = *lenp;
1807 if (len)
1808 if (copy_to_user(buffer, data, len))
1809 return -EFAULT;
1810 if (len < *lenp) {
1811 if (put_user('\n', buffer + len))
1812 return -EFAULT;
1813 len++;
1814 }
1815 *lenp = len;
1816 *ppos += len;
1817 }
1818 return 0;
1819 }
1820
1821 static void warn_sysctl_write(struct ctl_table *table)
1822 {
1823 pr_warn_once("%s wrote to %s when file position was not 0!\n"
1824 "This will not be supported in the future. To silence this\n"
1825 "warning, set kernel.sysctl_writes_strict = -1\n",
1826 current->comm, table->procname);
1827 }
1828
1829 /**
1830 * proc_dostring - read a string sysctl
1831 * @table: the sysctl table
1832 * @write: %TRUE if this is a write to the sysctl file
1833 * @buffer: the user buffer
1834 * @lenp: the size of the user buffer
1835 * @ppos: file position
1836 *
1837 * Reads/writes a string from/to the user buffer. If the kernel
1838 * buffer provided is not large enough to hold the string, the
1839 * string is truncated. The copied string is %NULL-terminated.
1840 * If the string is being read by the user process, it is copied
1841 * and a newline '\n' is added. It is truncated if the buffer is
1842 * not large enough.
1843 *
1844 * Returns 0 on success.
1845 */
1846 int proc_dostring(struct ctl_table *table, int write,
1847 void __user *buffer, size_t *lenp, loff_t *ppos)
1848 {
1849 if (write && *ppos && sysctl_writes_strict == SYSCTL_WRITES_WARN)
1850 warn_sysctl_write(table);
1851
1852 return _proc_do_string((char *)(table->data), table->maxlen, write,
1853 (char __user *)buffer, lenp, ppos);
1854 }
1855
1856 static size_t proc_skip_spaces(char **buf)
1857 {
1858 size_t ret;
1859 char *tmp = skip_spaces(*buf);
1860 ret = tmp - *buf;
1861 *buf = tmp;
1862 return ret;
1863 }
1864
1865 static void proc_skip_char(char **buf, size_t *size, const char v)
1866 {
1867 while (*size) {
1868 if (**buf != v)
1869 break;
1870 (*size)--;
1871 (*buf)++;
1872 }
1873 }
1874
1875 #define TMPBUFLEN 22
1876 /**
1877 * proc_get_long - reads an ASCII formatted integer from a user buffer
1878 *
1879 * @buf: a kernel buffer
1880 * @size: size of the kernel buffer
1881 * @val: this is where the number will be stored
1882 * @neg: set to %TRUE if number is negative
1883 * @perm_tr: a vector which contains the allowed trailers
1884 * @perm_tr_len: size of the perm_tr vector
1885 * @tr: pointer to store the trailer character
1886 *
1887 * In case of success %0 is returned and @buf and @size are updated with
1888 * the amount of bytes read. If @tr is non-NULL and a trailing
1889 * character exists (size is non-zero after returning from this
1890 * function), @tr is updated with the trailing character.
1891 */
1892 static int proc_get_long(char **buf, size_t *size,
1893 unsigned long *val, bool *neg,
1894 const char *perm_tr, unsigned perm_tr_len, char *tr)
1895 {
1896 int len;
1897 char *p, tmp[TMPBUFLEN];
1898
1899 if (!*size)
1900 return -EINVAL;
1901
1902 len = *size;
1903 if (len > TMPBUFLEN - 1)
1904 len = TMPBUFLEN - 1;
1905
1906 memcpy(tmp, *buf, len);
1907
1908 tmp[len] = 0;
1909 p = tmp;
1910 if (*p == '-' && *size > 1) {
1911 *neg = true;
1912 p++;
1913 } else
1914 *neg = false;
1915 if (!isdigit(*p))
1916 return -EINVAL;
1917
1918 *val = simple_strtoul(p, &p, 0);
1919
1920 len = p - tmp;
1921
1922 /* We don't know if the next char is whitespace thus we may accept
1923 * invalid integers (e.g. 1234...a) or two integers instead of one
1924 * (e.g. 123...1). So lets not allow such large numbers. */
1925 if (len == TMPBUFLEN - 1)
1926 return -EINVAL;
1927
1928 if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
1929 return -EINVAL;
1930
1931 if (tr && (len < *size))
1932 *tr = *p;
1933
1934 *buf += len;
1935 *size -= len;
1936
1937 return 0;
1938 }
1939
1940 /**
1941 * proc_put_long - converts an integer to a decimal ASCII formatted string
1942 *
1943 * @buf: the user buffer
1944 * @size: the size of the user buffer
1945 * @val: the integer to be converted
1946 * @neg: sign of the number, %TRUE for negative
1947 *
1948 * In case of success %0 is returned and @buf and @size are updated with
1949 * the amount of bytes written.
1950 */
1951 static int proc_put_long(void __user **buf, size_t *size, unsigned long val,
1952 bool neg)
1953 {
1954 int len;
1955 char tmp[TMPBUFLEN], *p = tmp;
1956
1957 sprintf(p, "%s%lu", neg ? "-" : "", val);
1958 len = strlen(tmp);
1959 if (len > *size)
1960 len = *size;
1961 if (copy_to_user(*buf, tmp, len))
1962 return -EFAULT;
1963 *size -= len;
1964 *buf += len;
1965 return 0;
1966 }
1967 #undef TMPBUFLEN
1968
1969 static int proc_put_char(void __user **buf, size_t *size, char c)
1970 {
1971 if (*size) {
1972 char __user **buffer = (char __user **)buf;
1973 if (put_user(c, *buffer))
1974 return -EFAULT;
1975 (*size)--, (*buffer)++;
1976 *buf = *buffer;
1977 }
1978 return 0;
1979 }
1980
1981 static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
1982 int *valp,
1983 int write, void *data)
1984 {
1985 if (write) {
1986 *valp = *negp ? -*lvalp : *lvalp;
1987 } else {
1988 int val = *valp;
1989 if (val < 0) {
1990 *negp = true;
1991 *lvalp = (unsigned long)-val;
1992 } else {
1993 *negp = false;
1994 *lvalp = (unsigned long)val;
1995 }
1996 }
1997 return 0;
1998 }
1999
2000 static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
2001
2002 static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
2003 int write, void __user *buffer,
2004 size_t *lenp, loff_t *ppos,
2005 int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
2006 int write, void *data),
2007 void *data)
2008 {
2009 int *i, vleft, first = 1, err = 0;
2010 unsigned long page = 0;
2011 size_t left;
2012 char *kbuf;
2013
2014 if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
2015 *lenp = 0;
2016 return 0;
2017 }
2018
2019 i = (int *) tbl_data;
2020 vleft = table->maxlen / sizeof(*i);
2021 left = *lenp;
2022
2023 if (!conv)
2024 conv = do_proc_dointvec_conv;
2025
2026 if (write) {
2027 if (*ppos) {
2028 switch (sysctl_writes_strict) {
2029 case SYSCTL_WRITES_STRICT:
2030 goto out;
2031 case SYSCTL_WRITES_WARN:
2032 warn_sysctl_write(table);
2033 break;
2034 default:
2035 break;
2036 }
2037 }
2038
2039 if (left > PAGE_SIZE - 1)
2040 left = PAGE_SIZE - 1;
2041 page = __get_free_page(GFP_TEMPORARY);
2042 kbuf = (char *) page;
2043 if (!kbuf)
2044 return -ENOMEM;
2045 if (copy_from_user(kbuf, buffer, left)) {
2046 err = -EFAULT;
2047 goto free;
2048 }
2049 kbuf[left] = 0;
2050 }
2051
2052 for (; left && vleft--; i++, first=0) {
2053 unsigned long lval;
2054 bool neg;
2055
2056 if (write) {
2057 left -= proc_skip_spaces(&kbuf);
2058
2059 if (!left)
2060 break;
2061 err = proc_get_long(&kbuf, &left, &lval, &neg,
2062 proc_wspace_sep,
2063 sizeof(proc_wspace_sep), NULL);
2064 if (err)
2065 break;
2066 if (conv(&neg, &lval, i, 1, data)) {
2067 err = -EINVAL;
2068 break;
2069 }
2070 } else {
2071 if (conv(&neg, &lval, i, 0, data)) {
2072 err = -EINVAL;
2073 break;
2074 }
2075 if (!first)
2076 err = proc_put_char(&buffer, &left, '\t');
2077 if (err)
2078 break;
2079 err = proc_put_long(&buffer, &left, lval, neg);
2080 if (err)
2081 break;
2082 }
2083 }
2084
2085 if (!write && !first && left && !err)
2086 err = proc_put_char(&buffer, &left, '\n');
2087 if (write && !err && left)
2088 left -= proc_skip_spaces(&kbuf);
2089 free:
2090 if (write) {
2091 free_page(page);
2092 if (first)
2093 return err ? : -EINVAL;
2094 }
2095 *lenp -= left;
2096 out:
2097 *ppos += *lenp;
2098 return err;
2099 }
2100
2101 static int do_proc_dointvec(struct ctl_table *table, int write,
2102 void __user *buffer, size_t *lenp, loff_t *ppos,
2103 int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
2104 int write, void *data),
2105 void *data)
2106 {
2107 return __do_proc_dointvec(table->data, table, write,
2108 buffer, lenp, ppos, conv, data);
2109 }
2110
2111 /**
2112 * proc_dointvec - read a vector of integers
2113 * @table: the sysctl table
2114 * @write: %TRUE if this is a write to the sysctl file
2115 * @buffer: the user buffer
2116 * @lenp: the size of the user buffer
2117 * @ppos: file position
2118 *
2119 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2120 * values from/to the user buffer, treated as an ASCII string.
2121 *
2122 * Returns 0 on success.
2123 */
2124 int proc_dointvec(struct ctl_table *table, int write,
2125 void __user *buffer, size_t *lenp, loff_t *ppos)
2126 {
2127 return do_proc_dointvec(table,write,buffer,lenp,ppos,
2128 NULL,NULL);
2129 }
2130
2131 /*
2132 * Taint values can only be increased
2133 * This means we can safely use a temporary.
2134 */
2135 static int proc_taint(struct ctl_table *table, int write,
2136 void __user *buffer, size_t *lenp, loff_t *ppos)
2137 {
2138 struct ctl_table t;
2139 unsigned long tmptaint = get_taint();
2140 int err;
2141
2142 if (write && !capable(CAP_SYS_ADMIN))
2143 return -EPERM;
2144
2145 t = *table;
2146 t.data = &tmptaint;
2147 err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
2148 if (err < 0)
2149 return err;
2150
2151 if (write) {
2152 /*
2153 * Poor man's atomic or. Not worth adding a primitive
2154 * to everyone's atomic.h for this
2155 */
2156 int i;
2157 for (i = 0; i < BITS_PER_LONG && tmptaint >> i; i++) {
2158 if ((tmptaint >> i) & 1)
2159 add_taint(i, LOCKDEP_STILL_OK);
2160 }
2161 }
2162
2163 return err;
2164 }
2165
2166 #ifdef CONFIG_PRINTK
2167 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
2168 void __user *buffer, size_t *lenp, loff_t *ppos)
2169 {
2170 if (write && !capable(CAP_SYS_ADMIN))
2171 return -EPERM;
2172
2173 return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2174 }
2175 #endif
2176
2177 struct do_proc_dointvec_minmax_conv_param {
2178 int *min;
2179 int *max;
2180 };
2181
2182 static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
2183 int *valp,
2184 int write, void *data)
2185 {
2186 struct do_proc_dointvec_minmax_conv_param *param = data;
2187 if (write) {
2188 int val = *negp ? -*lvalp : *lvalp;
2189 if ((param->min && *param->min > val) ||
2190 (param->max && *param->max < val))
2191 return -EINVAL;
2192 *valp = val;
2193 } else {
2194 int val = *valp;
2195 if (val < 0) {
2196 *negp = true;
2197 *lvalp = (unsigned long)-val;
2198 } else {
2199 *negp = false;
2200 *lvalp = (unsigned long)val;
2201 }
2202 }
2203 return 0;
2204 }
2205
2206 /**
2207 * proc_dointvec_minmax - read a vector of integers with min/max values
2208 * @table: the sysctl table
2209 * @write: %TRUE if this is a write to the sysctl file
2210 * @buffer: the user buffer
2211 * @lenp: the size of the user buffer
2212 * @ppos: file position
2213 *
2214 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2215 * values from/to the user buffer, treated as an ASCII string.
2216 *
2217 * This routine will ensure the values are within the range specified by
2218 * table->extra1 (min) and table->extra2 (max).
2219 *
2220 * Returns 0 on success.
2221 */
2222 int proc_dointvec_minmax(struct ctl_table *table, int write,
2223 void __user *buffer, size_t *lenp, loff_t *ppos)
2224 {
2225 struct do_proc_dointvec_minmax_conv_param param = {
2226 .min = (int *) table->extra1,
2227 .max = (int *) table->extra2,
2228 };
2229 return do_proc_dointvec(table, write, buffer, lenp, ppos,
2230 do_proc_dointvec_minmax_conv, &param);
2231 }
2232
2233 static void validate_coredump_safety(void)
2234 {
2235 #ifdef CONFIG_COREDUMP
2236 if (suid_dumpable == SUID_DUMP_ROOT &&
2237 core_pattern[0] != '/' && core_pattern[0] != '|') {
2238 printk(KERN_WARNING "Unsafe core_pattern used with "\
2239 "suid_dumpable=2. Pipe handler or fully qualified "\
2240 "core dump path required.\n");
2241 }
2242 #endif
2243 }
2244
2245 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
2246 void __user *buffer, size_t *lenp, loff_t *ppos)
2247 {
2248 int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2249 if (!error)
2250 validate_coredump_safety();
2251 return error;
2252 }
2253
2254 #ifdef CONFIG_COREDUMP
2255 static int proc_dostring_coredump(struct ctl_table *table, int write,
2256 void __user *buffer, size_t *lenp, loff_t *ppos)
2257 {
2258 int error = proc_dostring(table, write, buffer, lenp, ppos);
2259 if (!error)
2260 validate_coredump_safety();
2261 return error;
2262 }
2263 #endif
2264
2265 static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table, int write,
2266 void __user *buffer,
2267 size_t *lenp, loff_t *ppos,
2268 unsigned long convmul,
2269 unsigned long convdiv)
2270 {
2271 unsigned long *i, *min, *max;
2272 int vleft, first = 1, err = 0;
2273 unsigned long page = 0;
2274 size_t left;
2275 char *kbuf;
2276
2277 if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
2278 *lenp = 0;
2279 return 0;
2280 }
2281
2282 i = (unsigned long *) data;
2283 min = (unsigned long *) table->extra1;
2284 max = (unsigned long *) table->extra2;
2285 vleft = table->maxlen / sizeof(unsigned long);
2286 left = *lenp;
2287
2288 if (write) {
2289 if (*ppos) {
2290 switch (sysctl_writes_strict) {
2291 case SYSCTL_WRITES_STRICT:
2292 goto out;
2293 case SYSCTL_WRITES_WARN:
2294 warn_sysctl_write(table);
2295 break;
2296 default:
2297 break;
2298 }
2299 }
2300
2301 if (left > PAGE_SIZE - 1)
2302 left = PAGE_SIZE - 1;
2303 page = __get_free_page(GFP_TEMPORARY);
2304 kbuf = (char *) page;
2305 if (!kbuf)
2306 return -ENOMEM;
2307 if (copy_from_user(kbuf, buffer, left)) {
2308 err = -EFAULT;
2309 goto free;
2310 }
2311 kbuf[left] = 0;
2312 }
2313
2314 for (; left && vleft--; i++, first = 0) {
2315 unsigned long val;
2316
2317 if (write) {
2318 bool neg;
2319
2320 left -= proc_skip_spaces(&kbuf);
2321
2322 err = proc_get_long(&kbuf, &left, &val, &neg,
2323 proc_wspace_sep,
2324 sizeof(proc_wspace_sep), NULL);
2325 if (err)
2326 break;
2327 if (neg)
2328 continue;
2329 if ((min && val < *min) || (max && val > *max))
2330 continue;
2331 *i = val;
2332 } else {
2333 val = convdiv * (*i) / convmul;
2334 if (!first) {
2335 err = proc_put_char(&buffer, &left, '\t');
2336 if (err)
2337 break;
2338 }
2339 err = proc_put_long(&buffer, &left, val, false);
2340 if (err)
2341 break;
2342 }
2343 }
2344
2345 if (!write && !first && left && !err)
2346 err = proc_put_char(&buffer, &left, '\n');
2347 if (write && !err)
2348 left -= proc_skip_spaces(&kbuf);
2349 free:
2350 if (write) {
2351 free_page(page);
2352 if (first)
2353 return err ? : -EINVAL;
2354 }
2355 *lenp -= left;
2356 out:
2357 *ppos += *lenp;
2358 return err;
2359 }
2360
2361 static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
2362 void __user *buffer,
2363 size_t *lenp, loff_t *ppos,
2364 unsigned long convmul,
2365 unsigned long convdiv)
2366 {
2367 return __do_proc_doulongvec_minmax(table->data, table, write,
2368 buffer, lenp, ppos, convmul, convdiv);
2369 }
2370
2371 /**
2372 * proc_doulongvec_minmax - read a vector of long integers with min/max values
2373 * @table: the sysctl table
2374 * @write: %TRUE if this is a write to the sysctl file
2375 * @buffer: the user buffer
2376 * @lenp: the size of the user buffer
2377 * @ppos: file position
2378 *
2379 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2380 * values from/to the user buffer, treated as an ASCII string.
2381 *
2382 * This routine will ensure the values are within the range specified by
2383 * table->extra1 (min) and table->extra2 (max).
2384 *
2385 * Returns 0 on success.
2386 */
2387 int proc_doulongvec_minmax(struct ctl_table *table, int write,
2388 void __user *buffer, size_t *lenp, loff_t *ppos)
2389 {
2390 return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
2391 }
2392
2393 /**
2394 * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
2395 * @table: the sysctl table
2396 * @write: %TRUE if this is a write to the sysctl file
2397 * @buffer: the user buffer
2398 * @lenp: the size of the user buffer
2399 * @ppos: file position
2400 *
2401 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2402 * values from/to the user buffer, treated as an ASCII string. The values
2403 * are treated as milliseconds, and converted to jiffies when they are stored.
2404 *
2405 * This routine will ensure the values are within the range specified by
2406 * table->extra1 (min) and table->extra2 (max).
2407 *
2408 * Returns 0 on success.
2409 */
2410 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
2411 void __user *buffer,
2412 size_t *lenp, loff_t *ppos)
2413 {
2414 return do_proc_doulongvec_minmax(table, write, buffer,
2415 lenp, ppos, HZ, 1000l);
2416 }
2417
2418
2419 static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
2420 int *valp,
2421 int write, void *data)
2422 {
2423 if (write) {
2424 if (*lvalp > LONG_MAX / HZ)
2425 return 1;
2426 *valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
2427 } else {
2428 int val = *valp;
2429 unsigned long lval;
2430 if (val < 0) {
2431 *negp = true;
2432 lval = (unsigned long)-val;
2433 } else {
2434 *negp = false;
2435 lval = (unsigned long)val;
2436 }
2437 *lvalp = lval / HZ;
2438 }
2439 return 0;
2440 }
2441
2442 static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
2443 int *valp,
2444 int write, void *data)
2445 {
2446 if (write) {
2447 if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
2448 return 1;
2449 *valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
2450 } else {
2451 int val = *valp;
2452 unsigned long lval;
2453 if (val < 0) {
2454 *negp = true;
2455 lval = (unsigned long)-val;
2456 } else {
2457 *negp = false;
2458 lval = (unsigned long)val;
2459 }
2460 *lvalp = jiffies_to_clock_t(lval);
2461 }
2462 return 0;
2463 }
2464
2465 static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
2466 int *valp,
2467 int write, void *data)
2468 {
2469 if (write) {
2470 unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
2471
2472 if (jif > INT_MAX)
2473 return 1;
2474 *valp = (int)jif;
2475 } else {
2476 int val = *valp;
2477 unsigned long lval;
2478 if (val < 0) {
2479 *negp = true;
2480 lval = (unsigned long)-val;
2481 } else {
2482 *negp = false;
2483 lval = (unsigned long)val;
2484 }
2485 *lvalp = jiffies_to_msecs(lval);
2486 }
2487 return 0;
2488 }
2489
2490 /**
2491 * proc_dointvec_jiffies - read a vector of integers as seconds
2492 * @table: the sysctl table
2493 * @write: %TRUE if this is a write to the sysctl file
2494 * @buffer: the user buffer
2495 * @lenp: the size of the user buffer
2496 * @ppos: file position
2497 *
2498 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2499 * values from/to the user buffer, treated as an ASCII string.
2500 * The values read are assumed to be in seconds, and are converted into
2501 * jiffies.
2502 *
2503 * Returns 0 on success.
2504 */
2505 int proc_dointvec_jiffies(struct ctl_table *table, int write,
2506 void __user *buffer, size_t *lenp, loff_t *ppos)
2507 {
2508 return do_proc_dointvec(table,write,buffer,lenp,ppos,
2509 do_proc_dointvec_jiffies_conv,NULL);
2510 }
2511
2512 /**
2513 * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
2514 * @table: the sysctl table
2515 * @write: %TRUE if this is a write to the sysctl file
2516 * @buffer: the user buffer
2517 * @lenp: the size of the user buffer
2518 * @ppos: pointer to the file position
2519 *
2520 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2521 * values from/to the user buffer, treated as an ASCII string.
2522 * The values read are assumed to be in 1/USER_HZ seconds, and
2523 * are converted into jiffies.
2524 *
2525 * Returns 0 on success.
2526 */
2527 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
2528 void __user *buffer, size_t *lenp, loff_t *ppos)
2529 {
2530 return do_proc_dointvec(table,write,buffer,lenp,ppos,
2531 do_proc_dointvec_userhz_jiffies_conv,NULL);
2532 }
2533
2534 /**
2535 * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
2536 * @table: the sysctl table
2537 * @write: %TRUE if this is a write to the sysctl file
2538 * @buffer: the user buffer
2539 * @lenp: the size of the user buffer
2540 * @ppos: file position
2541 * @ppos: the current position in the file
2542 *
2543 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2544 * values from/to the user buffer, treated as an ASCII string.
2545 * The values read are assumed to be in 1/1000 seconds, and
2546 * are converted into jiffies.
2547 *
2548 * Returns 0 on success.
2549 */
2550 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
2551 void __user *buffer, size_t *lenp, loff_t *ppos)
2552 {
2553 return do_proc_dointvec(table, write, buffer, lenp, ppos,
2554 do_proc_dointvec_ms_jiffies_conv, NULL);
2555 }
2556
2557 static int proc_do_cad_pid(struct ctl_table *table, int write,
2558 void __user *buffer, size_t *lenp, loff_t *ppos)
2559 {
2560 struct pid *new_pid;
2561 pid_t tmp;
2562 int r;
2563
2564 tmp = pid_vnr(cad_pid);
2565
2566 r = __do_proc_dointvec(&tmp, table, write, buffer,
2567 lenp, ppos, NULL, NULL);
2568 if (r || !write)
2569 return r;
2570
2571 new_pid = find_get_pid(tmp);
2572 if (!new_pid)
2573 return -ESRCH;
2574
2575 put_pid(xchg(&cad_pid, new_pid));
2576 return 0;
2577 }
2578
2579 /**
2580 * proc_do_large_bitmap - read/write from/to a large bitmap
2581 * @table: the sysctl table
2582 * @write: %TRUE if this is a write to the sysctl file
2583 * @buffer: the user buffer
2584 * @lenp: the size of the user buffer
2585 * @ppos: file position
2586 *
2587 * The bitmap is stored at table->data and the bitmap length (in bits)
2588 * in table->maxlen.
2589 *
2590 * We use a range comma separated format (e.g. 1,3-4,10-10) so that
2591 * large bitmaps may be represented in a compact manner. Writing into
2592 * the file will clear the bitmap then update it with the given input.
2593 *
2594 * Returns 0 on success.
2595 */
2596 int proc_do_large_bitmap(struct ctl_table *table, int write,
2597 void __user *buffer, size_t *lenp, loff_t *ppos)
2598 {
2599 int err = 0;
2600 bool first = 1;
2601 size_t left = *lenp;
2602 unsigned long bitmap_len = table->maxlen;
2603 unsigned long *bitmap = *(unsigned long **) table->data;
2604 unsigned long *tmp_bitmap = NULL;
2605 char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
2606
2607 if (!bitmap || !bitmap_len || !left || (*ppos && !write)) {
2608 *lenp = 0;
2609 return 0;
2610 }
2611
2612 if (write) {
2613 unsigned long page = 0;
2614 char *kbuf;
2615
2616 if (left > PAGE_SIZE - 1)
2617 left = PAGE_SIZE - 1;
2618
2619 page = __get_free_page(GFP_TEMPORARY);
2620 kbuf = (char *) page;
2621 if (!kbuf)
2622 return -ENOMEM;
2623 if (copy_from_user(kbuf, buffer, left)) {
2624 free_page(page);
2625 return -EFAULT;
2626 }
2627 kbuf[left] = 0;
2628
2629 tmp_bitmap = kzalloc(BITS_TO_LONGS(bitmap_len) * sizeof(unsigned long),
2630 GFP_KERNEL);
2631 if (!tmp_bitmap) {
2632 free_page(page);
2633 return -ENOMEM;
2634 }
2635 proc_skip_char(&kbuf, &left, '\n');
2636 while (!err && left) {
2637 unsigned long val_a, val_b;
2638 bool neg;
2639
2640 err = proc_get_long(&kbuf, &left, &val_a, &neg, tr_a,
2641 sizeof(tr_a), &c);
2642 if (err)
2643 break;
2644 if (val_a >= bitmap_len || neg) {
2645 err = -EINVAL;
2646 break;
2647 }
2648
2649 val_b = val_a;
2650 if (left) {
2651 kbuf++;
2652 left--;
2653 }
2654
2655 if (c == '-') {
2656 err = proc_get_long(&kbuf, &left, &val_b,
2657 &neg, tr_b, sizeof(tr_b),
2658 &c);
2659 if (err)
2660 break;
2661 if (val_b >= bitmap_len || neg ||
2662 val_a > val_b) {
2663 err = -EINVAL;
2664 break;
2665 }
2666 if (left) {
2667 kbuf++;
2668 left--;
2669 }
2670 }
2671
2672 bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
2673 first = 0;
2674 proc_skip_char(&kbuf, &left, '\n');
2675 }
2676 free_page(page);
2677 } else {
2678 unsigned long bit_a, bit_b = 0;
2679
2680 while (left) {
2681 bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
2682 if (bit_a >= bitmap_len)
2683 break;
2684 bit_b = find_next_zero_bit(bitmap, bitmap_len,
2685 bit_a + 1) - 1;
2686
2687 if (!first) {
2688 err = proc_put_char(&buffer, &left, ',');
2689 if (err)
2690 break;
2691 }
2692 err = proc_put_long(&buffer, &left, bit_a, false);
2693 if (err)
2694 break;
2695 if (bit_a != bit_b) {
2696 err = proc_put_char(&buffer, &left, '-');
2697 if (err)
2698 break;
2699 err = proc_put_long(&buffer, &left, bit_b, false);
2700 if (err)
2701 break;
2702 }
2703
2704 first = 0; bit_b++;
2705 }
2706 if (!err)
2707 err = proc_put_char(&buffer, &left, '\n');
2708 }
2709
2710 if (!err) {
2711 if (write) {
2712 if (*ppos)
2713 bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
2714 else
2715 bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
2716 }
2717 kfree(tmp_bitmap);
2718 *lenp -= left;
2719 *ppos += *lenp;
2720 return 0;
2721 } else {
2722 kfree(tmp_bitmap);
2723 return err;
2724 }
2725 }
2726
2727 #else /* CONFIG_PROC_SYSCTL */
2728
2729 int proc_dostring(struct ctl_table *table, int write,
2730 void __user *buffer, size_t *lenp, loff_t *ppos)
2731 {
2732 return -ENOSYS;
2733 }
2734
2735 int proc_dointvec(struct ctl_table *table, int write,
2736 void __user *buffer, size_t *lenp, loff_t *ppos)
2737 {
2738 return -ENOSYS;
2739 }
2740
2741 int proc_dointvec_minmax(struct ctl_table *table, int write,
2742 void __user *buffer, size_t *lenp, loff_t *ppos)
2743 {
2744 return -ENOSYS;
2745 }
2746
2747 int proc_dointvec_jiffies(struct ctl_table *table, int write,
2748 void __user *buffer, size_t *lenp, loff_t *ppos)
2749 {
2750 return -ENOSYS;
2751 }
2752
2753 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
2754 void __user *buffer, size_t *lenp, loff_t *ppos)
2755 {
2756 return -ENOSYS;
2757 }
2758
2759 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
2760 void __user *buffer, size_t *lenp, loff_t *ppos)
2761 {
2762 return -ENOSYS;
2763 }
2764
2765 int proc_doulongvec_minmax(struct ctl_table *table, int write,
2766 void __user *buffer, size_t *lenp, loff_t *ppos)
2767 {
2768 return -ENOSYS;
2769 }
2770
2771 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
2772 void __user *buffer,
2773 size_t *lenp, loff_t *ppos)
2774 {
2775 return -ENOSYS;
2776 }
2777
2778
2779 #endif /* CONFIG_PROC_SYSCTL */
2780
2781 /*
2782 * No sense putting this after each symbol definition, twice,
2783 * exception granted :-)
2784 */
2785 EXPORT_SYMBOL(proc_dointvec);
2786 EXPORT_SYMBOL(proc_dointvec_jiffies);
2787 EXPORT_SYMBOL(proc_dointvec_minmax);
2788 EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
2789 EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
2790 EXPORT_SYMBOL(proc_dostring);
2791 EXPORT_SYMBOL(proc_doulongvec_minmax);
2792 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
Ubuntu Artful kernel mirror