?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
ARM: use read_cpuid_id() instead of read_cpuid(CPUID_ID)
[linux-2.6.git] / kernel / sysctl.c
blobc88878db491e44bd2b5a5ea7f381fc7242187727
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/mm.h>
23 #include <linux/swap.h>
24 #include <linux/slab.h>
25 #include <linux/sysctl.h>
26 #include <linux/bitmap.h>
27 #include <linux/signal.h>
28 #include <linux/printk.h>
29 #include <linux/proc_fs.h>
30 #include <linux/security.h>
31 #include <linux/ctype.h>
32 #include <linux/kmemcheck.h>
33 #include <linux/kmemleak.h>
34 #include <linux/fs.h>
35 #include <linux/init.h>
36 #include <linux/kernel.h>
37 #include <linux/kobject.h>
38 #include <linux/net.h>
39 #include <linux/sysrq.h>
40 #include <linux/highuid.h>
41 #include <linux/writeback.h>
42 #include <linux/ratelimit.h>
43 #include <linux/compaction.h>
44 #include <linux/hugetlb.h>
45 #include <linux/initrd.h>
46 #include <linux/key.h>
47 #include <linux/times.h>
48 #include <linux/limits.h>
49 #include <linux/dcache.h>
50 #include <linux/dnotify.h>
51 #include <linux/syscalls.h>
52 #include <linux/vmstat.h>
53 #include <linux/nfs_fs.h>
54 #include <linux/acpi.h>
55 #include <linux/reboot.h>
56 #include <linux/ftrace.h>
57 #include <linux/perf_event.h>
58 #include <linux/kprobes.h>
59 #include <linux/pipe_fs_i.h>
60 #include <linux/oom.h>
61 #include <linux/kmod.h>
62 #include <linux/capability.h>
63 #include <linux/binfmts.h>
64
65 #include <asm/uaccess.h>
66 #include <asm/processor.h>
67
68 #ifdef CONFIG_X86
69 #include <asm/nmi.h>
70 #include <asm/stacktrace.h>
71 #include <asm/io.h>
72 #endif
73 #ifdef CONFIG_SPARC
74 #include <asm/setup.h>
75 #endif
76 #ifdef CONFIG_BSD_PROCESS_ACCT
77 #include <linux/acct.h>
78 #endif
79 #ifdef CONFIG_RT_MUTEXES
80 #include <linux/rtmutex.h>
81 #endif
82 #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
83 #include <linux/lockdep.h>
84 #endif
85 #ifdef CONFIG_CHR_DEV_SG
86 #include <scsi/sg.h>
87 #endif
88
89 #ifdef CONFIG_LOCKUP_DETECTOR
90 #include <linux/nmi.h>
91 #endif
92
93
94 #if defined(CONFIG_SYSCTL)
95
96 /* External variables not in a header file. */
97 extern int sysctl_overcommit_memory;
98 extern int sysctl_overcommit_ratio;
99 extern int max_threads;
100 extern int suid_dumpable;
101 #ifdef CONFIG_COREDUMP
102 extern int core_uses_pid;
103 extern char core_pattern[];
104 extern unsigned int core_pipe_limit;
105 #endif
106 extern int pid_max;
107 extern int min_free_kbytes;
108 extern int pid_max_min, pid_max_max;
109 extern int sysctl_drop_caches;
110 extern int percpu_pagelist_fraction;
111 extern int compat_log;
112 extern int latencytop_enabled;
113 extern int sysctl_nr_open_min, sysctl_nr_open_max;
114 #ifndef CONFIG_MMU
115 extern int sysctl_nr_trim_pages;
116 #endif
117 #ifdef CONFIG_BLOCK
118 extern int blk_iopoll_enabled;
119 #endif
120
121 /* Constants used for minimum and maximum */
122 #ifdef CONFIG_LOCKUP_DETECTOR
123 static int sixty = 60;
124 static int neg_one = -1;
125 #endif
126
127 static int zero;
128 static int __maybe_unused one = 1;
129 static int __maybe_unused two = 2;
130 static int __maybe_unused three = 3;
131 static unsigned long one_ul = 1;
132 static int one_hundred = 100;
133 #ifdef CONFIG_PRINTK
134 static int ten_thousand = 10000;
135 #endif
136
137 /* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
138 static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
139
140 /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
141 static int maxolduid = 65535;
142 static int minolduid;
143 static int min_percpu_pagelist_fract = 8;
144
145 static int ngroups_max = NGROUPS_MAX;
146 static const int cap_last_cap = CAP_LAST_CAP;
147
148 #ifdef CONFIG_INOTIFY_USER
149 #include <linux/inotify.h>
150 #endif
151 #ifdef CONFIG_SPARC
152 #endif
153
154 #ifdef CONFIG_SPARC64
155 extern int sysctl_tsb_ratio;
156 #endif
157
158 #ifdef __hppa__
159 extern int pwrsw_enabled;
160 extern int unaligned_enabled;
161 #endif
162
163 #ifdef CONFIG_IA64
164 extern int no_unaligned_warning;
165 extern int unaligned_dump_stack;
166 #endif
167
168 #ifdef CONFIG_PROC_SYSCTL
169 static int proc_do_cad_pid(struct ctl_table *table, int write,
170 void __user *buffer, size_t *lenp, loff_t *ppos);
171 static int proc_taint(struct ctl_table *table, int write,
172 void __user *buffer, size_t *lenp, loff_t *ppos);
173 #endif
174
175 #ifdef CONFIG_PRINTK
176 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
177 void __user *buffer, size_t *lenp, loff_t *ppos);
178 #endif
179
180 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
181 void __user *buffer, size_t *lenp, loff_t *ppos);
182 #ifdef CONFIG_COREDUMP
183 static int proc_dostring_coredump(struct ctl_table *table, int write,
184 void __user *buffer, size_t *lenp, loff_t *ppos);
185 #endif
186
187 #ifdef CONFIG_MAGIC_SYSRQ
188 /* Note: sysrq code uses it's own private copy */
189 static int __sysrq_enabled = SYSRQ_DEFAULT_ENABLE;
190
191 static int sysrq_sysctl_handler(ctl_table *table, int write,
192 void __user *buffer, size_t *lenp,
193 loff_t *ppos)
194 {
195 int error;
196
197 error = proc_dointvec(table, write, buffer, lenp, ppos);
198 if (error)
199 return error;
200
201 if (write)
202 sysrq_toggle_support(__sysrq_enabled);
203
204 return 0;
205 }
206
207 #endif
208
209 static struct ctl_table kern_table[];
210 static struct ctl_table vm_table[];
211 static struct ctl_table fs_table[];
212 static struct ctl_table debug_table[];
213 static struct ctl_table dev_table[];
214 extern struct ctl_table random_table[];
215 #ifdef CONFIG_EPOLL
216 extern struct ctl_table epoll_table[];
217 #endif
218
219 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
220 int sysctl_legacy_va_layout;
221 #endif
222
223 /* The default sysctl tables: */
224
225 static struct ctl_table sysctl_base_table[] = {
226 {
227 .procname = "kernel",
228 .mode = 0555,
229 .child = kern_table,
230 },
231 {
232 .procname = "vm",
233 .mode = 0555,
234 .child = vm_table,
235 },
236 {
237 .procname = "fs",
238 .mode = 0555,
239 .child = fs_table,
240 },
241 {
242 .procname = "debug",
243 .mode = 0555,
244 .child = debug_table,
245 },
246 {
247 .procname = "dev",
248 .mode = 0555,
249 .child = dev_table,
250 },
251 { }
252 };
253
254 #ifdef CONFIG_SCHED_DEBUG
255 static int min_sched_granularity_ns = 100000; /* 100 usecs */
256 static int max_sched_granularity_ns = NSEC_PER_SEC; /* 1 second */
257 static int min_wakeup_granularity_ns; /* 0 usecs */
258 static int max_wakeup_granularity_ns = NSEC_PER_SEC; /* 1 second */
259 #ifdef CONFIG_SMP
260 static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
261 static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
262 #endif /* CONFIG_SMP */
263 #endif /* CONFIG_SCHED_DEBUG */
264
265 #ifdef CONFIG_COMPACTION
266 static int min_extfrag_threshold;
267 static int max_extfrag_threshold = 1000;
268 #endif
269
270 static struct ctl_table kern_table[] = {
271 {
272 .procname = "sched_child_runs_first",
273 .data = &sysctl_sched_child_runs_first,
274 .maxlen = sizeof(unsigned int),
275 .mode = 0644,
276 .proc_handler = proc_dointvec,
277 },
278 #ifdef CONFIG_SCHED_DEBUG
279 {
280 .procname = "sched_min_granularity_ns",
281 .data = &sysctl_sched_min_granularity,
282 .maxlen = sizeof(unsigned int),
283 .mode = 0644,
284 .proc_handler = sched_proc_update_handler,
285 .extra1 = &min_sched_granularity_ns,
286 .extra2 = &max_sched_granularity_ns,
287 },
288 {
289 .procname = "sched_latency_ns",
290 .data = &sysctl_sched_latency,
291 .maxlen = sizeof(unsigned int),
292 .mode = 0644,
293 .proc_handler = sched_proc_update_handler,
294 .extra1 = &min_sched_granularity_ns,
295 .extra2 = &max_sched_granularity_ns,
296 },
297 {
298 .procname = "sched_wakeup_granularity_ns",
299 .data = &sysctl_sched_wakeup_granularity,
300 .maxlen = sizeof(unsigned int),
301 .mode = 0644,
302 .proc_handler = sched_proc_update_handler,
303 .extra1 = &min_wakeup_granularity_ns,
304 .extra2 = &max_wakeup_granularity_ns,
305 },
306 #ifdef CONFIG_SMP
307 {
308 .procname = "sched_tunable_scaling",
309 .data = &sysctl_sched_tunable_scaling,
310 .maxlen = sizeof(enum sched_tunable_scaling),
311 .mode = 0644,
312 .proc_handler = sched_proc_update_handler,
313 .extra1 = &min_sched_tunable_scaling,
314 .extra2 = &max_sched_tunable_scaling,
315 },
316 {
317 .procname = "sched_migration_cost_ns",
318 .data = &sysctl_sched_migration_cost,
319 .maxlen = sizeof(unsigned int),
320 .mode = 0644,
321 .proc_handler = proc_dointvec,
322 },
323 {
324 .procname = "sched_nr_migrate",
325 .data = &sysctl_sched_nr_migrate,
326 .maxlen = sizeof(unsigned int),
327 .mode = 0644,
328 .proc_handler = proc_dointvec,
329 },
330 {
331 .procname = "sched_time_avg_ms",
332 .data = &sysctl_sched_time_avg,
333 .maxlen = sizeof(unsigned int),
334 .mode = 0644,
335 .proc_handler = proc_dointvec,
336 },
337 {
338 .procname = "sched_shares_window_ns",
339 .data = &sysctl_sched_shares_window,
340 .maxlen = sizeof(unsigned int),
341 .mode = 0644,
342 .proc_handler = proc_dointvec,
343 },
344 {
345 .procname = "timer_migration",
346 .data = &sysctl_timer_migration,
347 .maxlen = sizeof(unsigned int),
348 .mode = 0644,
349 .proc_handler = proc_dointvec_minmax,
350 .extra1 = &zero,
351 .extra2 = &one,
352 },
353 #endif /* CONFIG_SMP */
354 #ifdef CONFIG_NUMA_BALANCING
355 {
356 .procname = "numa_balancing_scan_delay_ms",
357 .data = &sysctl_numa_balancing_scan_delay,
358 .maxlen = sizeof(unsigned int),
359 .mode = 0644,
360 .proc_handler = proc_dointvec,
361 },
362 {
363 .procname = "numa_balancing_scan_period_min_ms",
364 .data = &sysctl_numa_balancing_scan_period_min,
365 .maxlen = sizeof(unsigned int),
366 .mode = 0644,
367 .proc_handler = proc_dointvec,
368 },
369 {
370 .procname = "numa_balancing_scan_period_reset",
371 .data = &sysctl_numa_balancing_scan_period_reset,
372 .maxlen = sizeof(unsigned int),
373 .mode = 0644,
374 .proc_handler = proc_dointvec,
375 },
376 {
377 .procname = "numa_balancing_scan_period_max_ms",
378 .data = &sysctl_numa_balancing_scan_period_max,
379 .maxlen = sizeof(unsigned int),
380 .mode = 0644,
381 .proc_handler = proc_dointvec,
382 },
383 {
384 .procname = "numa_balancing_scan_size_mb",
385 .data = &sysctl_numa_balancing_scan_size,
386 .maxlen = sizeof(unsigned int),
387 .mode = 0644,
388 .proc_handler = proc_dointvec,
389 },
390 #endif /* CONFIG_NUMA_BALANCING */
391 #endif /* CONFIG_SCHED_DEBUG */
392 {
393 .procname = "sched_rt_period_us",
394 .data = &sysctl_sched_rt_period,
395 .maxlen = sizeof(unsigned int),
396 .mode = 0644,
397 .proc_handler = sched_rt_handler,
398 },
399 {
400 .procname = "sched_rt_runtime_us",
401 .data = &sysctl_sched_rt_runtime,
402 .maxlen = sizeof(int),
403 .mode = 0644,
404 .proc_handler = sched_rt_handler,
405 },
406 #ifdef CONFIG_SCHED_AUTOGROUP
407 {
408 .procname = "sched_autogroup_enabled",
409 .data = &sysctl_sched_autogroup_enabled,
410 .maxlen = sizeof(unsigned int),
411 .mode = 0644,
412 .proc_handler = proc_dointvec_minmax,
413 .extra1 = &zero,
414 .extra2 = &one,
415 },
416 #endif
417 #ifdef CONFIG_CFS_BANDWIDTH
418 {
419 .procname = "sched_cfs_bandwidth_slice_us",
420 .data = &sysctl_sched_cfs_bandwidth_slice,
421 .maxlen = sizeof(unsigned int),
422 .mode = 0644,
423 .proc_handler = proc_dointvec_minmax,
424 .extra1 = &one,
425 },
426 #endif
427 #ifdef CONFIG_PROVE_LOCKING
428 {
429 .procname = "prove_locking",
430 .data = &prove_locking,
431 .maxlen = sizeof(int),
432 .mode = 0644,
433 .proc_handler = proc_dointvec,
434 },
435 #endif
436 #ifdef CONFIG_LOCK_STAT
437 {
438 .procname = "lock_stat",
439 .data = &lock_stat,
440 .maxlen = sizeof(int),
441 .mode = 0644,
442 .proc_handler = proc_dointvec,
443 },
444 #endif
445 {
446 .procname = "panic",
447 .data = &panic_timeout,
448 .maxlen = sizeof(int),
449 .mode = 0644,
450 .proc_handler = proc_dointvec,
451 },
452 #ifdef CONFIG_COREDUMP
453 {
454 .procname = "core_uses_pid",
455 .data = &core_uses_pid,
456 .maxlen = sizeof(int),
457 .mode = 0644,
458 .proc_handler = proc_dointvec,
459 },
460 {
461 .procname = "core_pattern",
462 .data = core_pattern,
463 .maxlen = CORENAME_MAX_SIZE,
464 .mode = 0644,
465 .proc_handler = proc_dostring_coredump,
466 },
467 {
468 .procname = "core_pipe_limit",
469 .data = &core_pipe_limit,
470 .maxlen = sizeof(unsigned int),
471 .mode = 0644,
472 .proc_handler = proc_dointvec,
473 },
474 #endif
475 #ifdef CONFIG_PROC_SYSCTL
476 {
477 .procname = "tainted",
478 .maxlen = sizeof(long),
479 .mode = 0644,
480 .proc_handler = proc_taint,
481 },
482 #endif
483 #ifdef CONFIG_LATENCYTOP
484 {
485 .procname = "latencytop",
486 .data = &latencytop_enabled,
487 .maxlen = sizeof(int),
488 .mode = 0644,
489 .proc_handler = proc_dointvec,
490 },
491 #endif
492 #ifdef CONFIG_BLK_DEV_INITRD
493 {
494 .procname = "real-root-dev",
495 .data = &real_root_dev,
496 .maxlen = sizeof(int),
497 .mode = 0644,
498 .proc_handler = proc_dointvec,
499 },
500 #endif
501 {
502 .procname = "print-fatal-signals",
503 .data = &print_fatal_signals,
504 .maxlen = sizeof(int),
505 .mode = 0644,
506 .proc_handler = proc_dointvec,
507 },
508 #ifdef CONFIG_SPARC
509 {
510 .procname = "reboot-cmd",
511 .data = reboot_command,
512 .maxlen = 256,
513 .mode = 0644,
514 .proc_handler = proc_dostring,
515 },
516 {
517 .procname = "stop-a",
518 .data = &stop_a_enabled,
519 .maxlen = sizeof (int),
520 .mode = 0644,
521 .proc_handler = proc_dointvec,
522 },
523 {
524 .procname = "scons-poweroff",
525 .data = &scons_pwroff,
526 .maxlen = sizeof (int),
527 .mode = 0644,
528 .proc_handler = proc_dointvec,
529 },
530 #endif
531 #ifdef CONFIG_SPARC64
532 {
533 .procname = "tsb-ratio",
534 .data = &sysctl_tsb_ratio,
535 .maxlen = sizeof (int),
536 .mode = 0644,
537 .proc_handler = proc_dointvec,
538 },
539 #endif
540 #ifdef __hppa__
541 {
542 .procname = "soft-power",
543 .data = &pwrsw_enabled,
544 .maxlen = sizeof (int),
545 .mode = 0644,
546 .proc_handler = proc_dointvec,
547 },
548 {
549 .procname = "unaligned-trap",
550 .data = &unaligned_enabled,
551 .maxlen = sizeof (int),
552 .mode = 0644,
553 .proc_handler = proc_dointvec,
554 },
555 #endif
556 {
557 .procname = "ctrl-alt-del",
558 .data = &C_A_D,
559 .maxlen = sizeof(int),
560 .mode = 0644,
561 .proc_handler = proc_dointvec,
562 },
563 #ifdef CONFIG_FUNCTION_TRACER
564 {
565 .procname = "ftrace_enabled",
566 .data = &ftrace_enabled,
567 .maxlen = sizeof(int),
568 .mode = 0644,
569 .proc_handler = ftrace_enable_sysctl,
570 },
571 #endif
572 #ifdef CONFIG_STACK_TRACER
573 {
574 .procname = "stack_tracer_enabled",
575 .data = &stack_tracer_enabled,
576 .maxlen = sizeof(int),
577 .mode = 0644,
578 .proc_handler = stack_trace_sysctl,
579 },
580 #endif
581 #ifdef CONFIG_TRACING
582 {
583 .procname = "ftrace_dump_on_oops",
584 .data = &ftrace_dump_on_oops,
585 .maxlen = sizeof(int),
586 .mode = 0644,
587 .proc_handler = proc_dointvec,
588 },
589 #endif
590 #ifdef CONFIG_MODULES
591 {
592 .procname = "modprobe",
593 .data = &modprobe_path,
594 .maxlen = KMOD_PATH_LEN,
595 .mode = 0644,
596 .proc_handler = proc_dostring,
597 },
598 {
599 .procname = "modules_disabled",
600 .data = &modules_disabled,
601 .maxlen = sizeof(int),
602 .mode = 0644,
603 /* only handle a transition from default "0" to "1" */
604 .proc_handler = proc_dointvec_minmax,
605 .extra1 = &one,
606 .extra2 = &one,
607 },
608 #endif
609
610 {
611 .procname = "hotplug",
612 .data = &uevent_helper,
613 .maxlen = UEVENT_HELPER_PATH_LEN,
614 .mode = 0644,
615 .proc_handler = proc_dostring,
616 },
617
618 #ifdef CONFIG_CHR_DEV_SG
619 {
620 .procname = "sg-big-buff",
621 .data = &sg_big_buff,
622 .maxlen = sizeof (int),
623 .mode = 0444,
624 .proc_handler = proc_dointvec,
625 },
626 #endif
627 #ifdef CONFIG_BSD_PROCESS_ACCT
628 {
629 .procname = "acct",
630 .data = &acct_parm,
631 .maxlen = 3*sizeof(int),
632 .mode = 0644,
633 .proc_handler = proc_dointvec,
634 },
635 #endif
636 #ifdef CONFIG_MAGIC_SYSRQ
637 {
638 .procname = "sysrq",
639 .data = &__sysrq_enabled,
640 .maxlen = sizeof (int),
641 .mode = 0644,
642 .proc_handler = sysrq_sysctl_handler,
643 },
644 #endif
645 #ifdef CONFIG_PROC_SYSCTL
646 {
647 .procname = "cad_pid",
648 .data = NULL,
649 .maxlen = sizeof (int),
650 .mode = 0600,
651 .proc_handler = proc_do_cad_pid,
652 },
653 #endif
654 {
655 .procname = "threads-max",
656 .data = &max_threads,
657 .maxlen = sizeof(int),
658 .mode = 0644,
659 .proc_handler = proc_dointvec,
660 },
661 {
662 .procname = "random",
663 .mode = 0555,
664 .child = random_table,
665 },
666 {
667 .procname = "usermodehelper",
668 .mode = 0555,
669 .child = usermodehelper_table,
670 },
671 {
672 .procname = "overflowuid",
673 .data = &overflowuid,
674 .maxlen = sizeof(int),
675 .mode = 0644,
676 .proc_handler = proc_dointvec_minmax,
677 .extra1 = &minolduid,
678 .extra2 = &maxolduid,
679 },
680 {
681 .procname = "overflowgid",
682 .data = &overflowgid,
683 .maxlen = sizeof(int),
684 .mode = 0644,
685 .proc_handler = proc_dointvec_minmax,
686 .extra1 = &minolduid,
687 .extra2 = &maxolduid,
688 },
689 #ifdef CONFIG_S390
690 #ifdef CONFIG_MATHEMU
691 {
692 .procname = "ieee_emulation_warnings",
693 .data = &sysctl_ieee_emulation_warnings,
694 .maxlen = sizeof(int),
695 .mode = 0644,
696 .proc_handler = proc_dointvec,
697 },
698 #endif
699 {
700 .procname = "userprocess_debug",
701 .data = &show_unhandled_signals,
702 .maxlen = sizeof(int),
703 .mode = 0644,
704 .proc_handler = proc_dointvec,
705 },
706 #endif
707 {
708 .procname = "pid_max",
709 .data = &pid_max,
710 .maxlen = sizeof (int),
711 .mode = 0644,
712 .proc_handler = proc_dointvec_minmax,
713 .extra1 = &pid_max_min,
714 .extra2 = &pid_max_max,
715 },
716 {
717 .procname = "panic_on_oops",
718 .data = &panic_on_oops,
719 .maxlen = sizeof(int),
720 .mode = 0644,
721 .proc_handler = proc_dointvec,
722 },
723 #if defined CONFIG_PRINTK
724 {
725 .procname = "printk",
726 .data = &console_loglevel,
727 .maxlen = 4*sizeof(int),
728 .mode = 0644,
729 .proc_handler = proc_dointvec,
730 },
731 {
732 .procname = "printk_ratelimit",
733 .data = &printk_ratelimit_state.interval,
734 .maxlen = sizeof(int),
735 .mode = 0644,
736 .proc_handler = proc_dointvec_jiffies,
737 },
738 {
739 .procname = "printk_ratelimit_burst",
740 .data = &printk_ratelimit_state.burst,
741 .maxlen = sizeof(int),
742 .mode = 0644,
743 .proc_handler = proc_dointvec,
744 },
745 {
746 .procname = "printk_delay",
747 .data = &printk_delay_msec,
748 .maxlen = sizeof(int),
749 .mode = 0644,
750 .proc_handler = proc_dointvec_minmax,
751 .extra1 = &zero,
752 .extra2 = &ten_thousand,
753 },
754 {
755 .procname = "dmesg_restrict",
756 .data = &dmesg_restrict,
757 .maxlen = sizeof(int),
758 .mode = 0644,
759 .proc_handler = proc_dointvec_minmax_sysadmin,
760 .extra1 = &zero,
761 .extra2 = &one,
762 },
763 {
764 .procname = "kptr_restrict",
765 .data = &kptr_restrict,
766 .maxlen = sizeof(int),
767 .mode = 0644,
768 .proc_handler = proc_dointvec_minmax_sysadmin,
769 .extra1 = &zero,
770 .extra2 = &two,
771 },
772 #endif
773 {
774 .procname = "ngroups_max",
775 .data = &ngroups_max,
776 .maxlen = sizeof (int),
777 .mode = 0444,
778 .proc_handler = proc_dointvec,
779 },
780 {
781 .procname = "cap_last_cap",
782 .data = (void *)&cap_last_cap,
783 .maxlen = sizeof(int),
784 .mode = 0444,
785 .proc_handler = proc_dointvec,
786 },
787 #if defined(CONFIG_LOCKUP_DETECTOR)
788 {
789 .procname = "watchdog",
790 .data = &watchdog_enabled,
791 .maxlen = sizeof (int),
792 .mode = 0644,
793 .proc_handler = proc_dowatchdog,
794 .extra1 = &zero,
795 .extra2 = &one,
796 },
797 {
798 .procname = "watchdog_thresh",
799 .data = &watchdog_thresh,
800 .maxlen = sizeof(int),
801 .mode = 0644,
802 .proc_handler = proc_dowatchdog,
803 .extra1 = &neg_one,
804 .extra2 = &sixty,
805 },
806 {
807 .procname = "softlockup_panic",
808 .data = &softlockup_panic,
809 .maxlen = sizeof(int),
810 .mode = 0644,
811 .proc_handler = proc_dointvec_minmax,
812 .extra1 = &zero,
813 .extra2 = &one,
814 },
815 {
816 .procname = "nmi_watchdog",
817 .data = &watchdog_enabled,
818 .maxlen = sizeof (int),
819 .mode = 0644,
820 .proc_handler = proc_dowatchdog,
821 .extra1 = &zero,
822 .extra2 = &one,
823 },
824 #endif
825 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
826 {
827 .procname = "unknown_nmi_panic",
828 .data = &unknown_nmi_panic,
829 .maxlen = sizeof (int),
830 .mode = 0644,
831 .proc_handler = proc_dointvec,
832 },
833 #endif
834 #if defined(CONFIG_X86)
835 {
836 .procname = "panic_on_unrecovered_nmi",
837 .data = &panic_on_unrecovered_nmi,
838 .maxlen = sizeof(int),
839 .mode = 0644,
840 .proc_handler = proc_dointvec,
841 },
842 {
843 .procname = "panic_on_io_nmi",
844 .data = &panic_on_io_nmi,
845 .maxlen = sizeof(int),
846 .mode = 0644,
847 .proc_handler = proc_dointvec,
848 },
849 #ifdef CONFIG_DEBUG_STACKOVERFLOW
850 {
851 .procname = "panic_on_stackoverflow",
852 .data = &sysctl_panic_on_stackoverflow,
853 .maxlen = sizeof(int),
854 .mode = 0644,
855 .proc_handler = proc_dointvec,
856 },
857 #endif
858 {
859 .procname = "bootloader_type",
860 .data = &bootloader_type,
861 .maxlen = sizeof (int),
862 .mode = 0444,
863 .proc_handler = proc_dointvec,
864 },
865 {
866 .procname = "bootloader_version",
867 .data = &bootloader_version,
868 .maxlen = sizeof (int),
869 .mode = 0444,
870 .proc_handler = proc_dointvec,
871 },
872 {
873 .procname = "kstack_depth_to_print",
874 .data = &kstack_depth_to_print,
875 .maxlen = sizeof(int),
876 .mode = 0644,
877 .proc_handler = proc_dointvec,
878 },
879 {
880 .procname = "io_delay_type",
881 .data = &io_delay_type,
882 .maxlen = sizeof(int),
883 .mode = 0644,
884 .proc_handler = proc_dointvec,
885 },
886 #endif
887 #if defined(CONFIG_MMU)
888 {
889 .procname = "randomize_va_space",
890 .data = &randomize_va_space,
891 .maxlen = sizeof(int),
892 .mode = 0644,
893 .proc_handler = proc_dointvec,
894 },
895 #endif
896 #if defined(CONFIG_S390) && defined(CONFIG_SMP)
897 {
898 .procname = "spin_retry",
899 .data = &spin_retry,
900 .maxlen = sizeof (int),
901 .mode = 0644,
902 .proc_handler = proc_dointvec,
903 },
904 #endif
905 #if defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
906 {
907 .procname = "acpi_video_flags",
908 .data = &acpi_realmode_flags,
909 .maxlen = sizeof (unsigned long),
910 .mode = 0644,
911 .proc_handler = proc_doulongvec_minmax,
912 },
913 #endif
914 #ifdef CONFIG_IA64
915 {
916 .procname = "ignore-unaligned-usertrap",
917 .data = &no_unaligned_warning,
918 .maxlen = sizeof (int),
919 .mode = 0644,
920 .proc_handler = proc_dointvec,
921 },
922 {
923 .procname = "unaligned-dump-stack",
924 .data = &unaligned_dump_stack,
925 .maxlen = sizeof (int),
926 .mode = 0644,
927 .proc_handler = proc_dointvec,
928 },
929 #endif
930 #ifdef CONFIG_DETECT_HUNG_TASK
931 {
932 .procname = "hung_task_panic",
933 .data = &sysctl_hung_task_panic,
934 .maxlen = sizeof(int),
935 .mode = 0644,
936 .proc_handler = proc_dointvec_minmax,
937 .extra1 = &zero,
938 .extra2 = &one,
939 },
940 {
941 .procname = "hung_task_check_count",
942 .data = &sysctl_hung_task_check_count,
943 .maxlen = sizeof(unsigned long),
944 .mode = 0644,
945 .proc_handler = proc_doulongvec_minmax,
946 },
947 {
948 .procname = "hung_task_timeout_secs",
949 .data = &sysctl_hung_task_timeout_secs,
950 .maxlen = sizeof(unsigned long),
951 .mode = 0644,
952 .proc_handler = proc_dohung_task_timeout_secs,
953 },
954 {
955 .procname = "hung_task_warnings",
956 .data = &sysctl_hung_task_warnings,
957 .maxlen = sizeof(unsigned long),
958 .mode = 0644,
959 .proc_handler = proc_doulongvec_minmax,
960 },
961 #endif
962 #ifdef CONFIG_COMPAT
963 {
964 .procname = "compat-log",
965 .data = &compat_log,
966 .maxlen = sizeof (int),
967 .mode = 0644,
968 .proc_handler = proc_dointvec,
969 },
970 #endif
971 #ifdef CONFIG_RT_MUTEXES
972 {
973 .procname = "max_lock_depth",
974 .data = &max_lock_depth,
975 .maxlen = sizeof(int),
976 .mode = 0644,
977 .proc_handler = proc_dointvec,
978 },
979 #endif
980 {
981 .procname = "poweroff_cmd",
982 .data = &poweroff_cmd,
983 .maxlen = POWEROFF_CMD_PATH_LEN,
984 .mode = 0644,
985 .proc_handler = proc_dostring,
986 },
987 #ifdef CONFIG_KEYS
988 {
989 .procname = "keys",
990 .mode = 0555,
991 .child = key_sysctls,
992 },
993 #endif
994 #ifdef CONFIG_RCU_TORTURE_TEST
995 {
996 .procname = "rcutorture_runnable",
997 .data = &rcutorture_runnable,
998 .maxlen = sizeof(int),
999 .mode = 0644,
1000 .proc_handler = proc_dointvec,
1001 },
1002 #endif
1003 #ifdef CONFIG_PERF_EVENTS
1004 /*
1005 * User-space scripts rely on the existence of this file
1006 * as a feature check for perf_events being enabled.
1007 *
1008 * So it's an ABI, do not remove!
1009 */
1010 {
1011 .procname = "perf_event_paranoid",
1012 .data = &sysctl_perf_event_paranoid,
1013 .maxlen = sizeof(sysctl_perf_event_paranoid),
1014 .mode = 0644,
1015 .proc_handler = proc_dointvec,
1016 },
1017 {
1018 .procname = "perf_event_mlock_kb",
1019 .data = &sysctl_perf_event_mlock,
1020 .maxlen = sizeof(sysctl_perf_event_mlock),
1021 .mode = 0644,
1022 .proc_handler = proc_dointvec,
1023 },
1024 {
1025 .procname = "perf_event_max_sample_rate",
1026 .data = &sysctl_perf_event_sample_rate,
1027 .maxlen = sizeof(sysctl_perf_event_sample_rate),
1028 .mode = 0644,
1029 .proc_handler = perf_proc_update_handler,
1030 },
1031 #endif
1032 #ifdef CONFIG_KMEMCHECK
1033 {
1034 .procname = "kmemcheck",
1035 .data = &kmemcheck_enabled,
1036 .maxlen = sizeof(int),
1037 .mode = 0644,
1038 .proc_handler = proc_dointvec,
1039 },
1040 #endif
1041 #ifdef CONFIG_BLOCK
1042 {
1043 .procname = "blk_iopoll",
1044 .data = &blk_iopoll_enabled,
1045 .maxlen = sizeof(int),
1046 .mode = 0644,
1047 .proc_handler = proc_dointvec,
1048 },
1049 #endif
1050 { }
1051 };
1052
1053 static struct ctl_table vm_table[] = {
1054 {
1055 .procname = "overcommit_memory",
1056 .data = &sysctl_overcommit_memory,
1057 .maxlen = sizeof(sysctl_overcommit_memory),
1058 .mode = 0644,
1059 .proc_handler = proc_dointvec_minmax,
1060 .extra1 = &zero,
1061 .extra2 = &two,
1062 },
1063 {
1064 .procname = "panic_on_oom",
1065 .data = &sysctl_panic_on_oom,
1066 .maxlen = sizeof(sysctl_panic_on_oom),
1067 .mode = 0644,
1068 .proc_handler = proc_dointvec_minmax,
1069 .extra1 = &zero,
1070 .extra2 = &two,
1071 },
1072 {
1073 .procname = "oom_kill_allocating_task",
1074 .data = &sysctl_oom_kill_allocating_task,
1075 .maxlen = sizeof(sysctl_oom_kill_allocating_task),
1076 .mode = 0644,
1077 .proc_handler = proc_dointvec,
1078 },
1079 {
1080 .procname = "oom_dump_tasks",
1081 .data = &sysctl_oom_dump_tasks,
1082 .maxlen = sizeof(sysctl_oom_dump_tasks),
1083 .mode = 0644,
1084 .proc_handler = proc_dointvec,
1085 },
1086 {
1087 .procname = "overcommit_ratio",
1088 .data = &sysctl_overcommit_ratio,
1089 .maxlen = sizeof(sysctl_overcommit_ratio),
1090 .mode = 0644,
1091 .proc_handler = proc_dointvec,
1092 },
1093 {
1094 .procname = "page-cluster",
1095 .data = &page_cluster,
1096 .maxlen = sizeof(int),
1097 .mode = 0644,
1098 .proc_handler = proc_dointvec_minmax,
1099 .extra1 = &zero,
1100 },
1101 {
1102 .procname = "dirty_background_ratio",
1103 .data = &dirty_background_ratio,
1104 .maxlen = sizeof(dirty_background_ratio),
1105 .mode = 0644,
1106 .proc_handler = dirty_background_ratio_handler,
1107 .extra1 = &zero,
1108 .extra2 = &one_hundred,
1109 },
1110 {
1111 .procname = "dirty_background_bytes",
1112 .data = &dirty_background_bytes,
1113 .maxlen = sizeof(dirty_background_bytes),
1114 .mode = 0644,
1115 .proc_handler = dirty_background_bytes_handler,
1116 .extra1 = &one_ul,
1117 },
1118 {
1119 .procname = "dirty_ratio",
1120 .data = &vm_dirty_ratio,
1121 .maxlen = sizeof(vm_dirty_ratio),
1122 .mode = 0644,
1123 .proc_handler = dirty_ratio_handler,
1124 .extra1 = &zero,
1125 .extra2 = &one_hundred,
1126 },
1127 {
1128 .procname = "dirty_bytes",
1129 .data = &vm_dirty_bytes,
1130 .maxlen = sizeof(vm_dirty_bytes),
1131 .mode = 0644,
1132 .proc_handler = dirty_bytes_handler,
1133 .extra1 = &dirty_bytes_min,
1134 },
1135 {
1136 .procname = "dirty_writeback_centisecs",
1137 .data = &dirty_writeback_interval,
1138 .maxlen = sizeof(dirty_writeback_interval),
1139 .mode = 0644,
1140 .proc_handler = dirty_writeback_centisecs_handler,
1141 },
1142 {
1143 .procname = "dirty_expire_centisecs",
1144 .data = &dirty_expire_interval,
1145 .maxlen = sizeof(dirty_expire_interval),
1146 .mode = 0644,
1147 .proc_handler = proc_dointvec_minmax,
1148 .extra1 = &zero,
1149 },
1150 {
1151 .procname = "nr_pdflush_threads",
1152 .mode = 0444 /* read-only */,
1153 .proc_handler = pdflush_proc_obsolete,
1154 },
1155 {
1156 .procname = "swappiness",
1157 .data = &vm_swappiness,
1158 .maxlen = sizeof(vm_swappiness),
1159 .mode = 0644,
1160 .proc_handler = proc_dointvec_minmax,
1161 .extra1 = &zero,
1162 .extra2 = &one_hundred,
1163 },
1164 #ifdef CONFIG_HUGETLB_PAGE
1165 {
1166 .procname = "nr_hugepages",
1167 .data = NULL,
1168 .maxlen = sizeof(unsigned long),
1169 .mode = 0644,
1170 .proc_handler = hugetlb_sysctl_handler,
1171 .extra1 = (void *)&hugetlb_zero,
1172 .extra2 = (void *)&hugetlb_infinity,
1173 },
1174 #ifdef CONFIG_NUMA
1175 {
1176 .procname = "nr_hugepages_mempolicy",
1177 .data = NULL,
1178 .maxlen = sizeof(unsigned long),
1179 .mode = 0644,
1180 .proc_handler = &hugetlb_mempolicy_sysctl_handler,
1181 .extra1 = (void *)&hugetlb_zero,
1182 .extra2 = (void *)&hugetlb_infinity,
1183 },
1184 #endif
1185 {
1186 .procname = "hugetlb_shm_group",
1187 .data = &sysctl_hugetlb_shm_group,
1188 .maxlen = sizeof(gid_t),
1189 .mode = 0644,
1190 .proc_handler = proc_dointvec,
1191 },
1192 {
1193 .procname = "hugepages_treat_as_movable",
1194 .data = &hugepages_treat_as_movable,
1195 .maxlen = sizeof(int),
1196 .mode = 0644,
1197 .proc_handler = hugetlb_treat_movable_handler,
1198 },
1199 {
1200 .procname = "nr_overcommit_hugepages",
1201 .data = NULL,
1202 .maxlen = sizeof(unsigned long),
1203 .mode = 0644,
1204 .proc_handler = hugetlb_overcommit_handler,
1205 .extra1 = (void *)&hugetlb_zero,
1206 .extra2 = (void *)&hugetlb_infinity,
1207 },
1208 #endif
1209 {
1210 .procname = "lowmem_reserve_ratio",
1211 .data = &sysctl_lowmem_reserve_ratio,
1212 .maxlen = sizeof(sysctl_lowmem_reserve_ratio),
1213 .mode = 0644,
1214 .proc_handler = lowmem_reserve_ratio_sysctl_handler,
1215 },
1216 {
1217 .procname = "drop_caches",
1218 .data = &sysctl_drop_caches,
1219 .maxlen = sizeof(int),
1220 .mode = 0644,
1221 .proc_handler = drop_caches_sysctl_handler,
1222 .extra1 = &one,
1223 .extra2 = &three,
1224 },
1225 #ifdef CONFIG_COMPACTION
1226 {
1227 .procname = "compact_memory",
1228 .data = &sysctl_compact_memory,
1229 .maxlen = sizeof(int),
1230 .mode = 0200,
1231 .proc_handler = sysctl_compaction_handler,
1232 },
1233 {
1234 .procname = "extfrag_threshold",
1235 .data = &sysctl_extfrag_threshold,
1236 .maxlen = sizeof(int),
1237 .mode = 0644,
1238 .proc_handler = sysctl_extfrag_handler,
1239 .extra1 = &min_extfrag_threshold,
1240 .extra2 = &max_extfrag_threshold,
1241 },
1242
1243 #endif /* CONFIG_COMPACTION */
1244 {
1245 .procname = "min_free_kbytes",
1246 .data = &min_free_kbytes,
1247 .maxlen = sizeof(min_free_kbytes),
1248 .mode = 0644,
1249 .proc_handler = min_free_kbytes_sysctl_handler,
1250 .extra1 = &zero,
1251 },
1252 {
1253 .procname = "percpu_pagelist_fraction",
1254 .data = &percpu_pagelist_fraction,
1255 .maxlen = sizeof(percpu_pagelist_fraction),
1256 .mode = 0644,
1257 .proc_handler = percpu_pagelist_fraction_sysctl_handler,
1258 .extra1 = &min_percpu_pagelist_fract,
1259 },
1260 #ifdef CONFIG_MMU
1261 {
1262 .procname = "max_map_count",
1263 .data = &sysctl_max_map_count,
1264 .maxlen = sizeof(sysctl_max_map_count),
1265 .mode = 0644,
1266 .proc_handler = proc_dointvec_minmax,
1267 .extra1 = &zero,
1268 },
1269 #else
1270 {
1271 .procname = "nr_trim_pages",
1272 .data = &sysctl_nr_trim_pages,
1273 .maxlen = sizeof(sysctl_nr_trim_pages),
1274 .mode = 0644,
1275 .proc_handler = proc_dointvec_minmax,
1276 .extra1 = &zero,
1277 },
1278 #endif
1279 {
1280 .procname = "laptop_mode",
1281 .data = &laptop_mode,
1282 .maxlen = sizeof(laptop_mode),
1283 .mode = 0644,
1284 .proc_handler = proc_dointvec_jiffies,
1285 },
1286 {
1287 .procname = "block_dump",
1288 .data = &block_dump,
1289 .maxlen = sizeof(block_dump),
1290 .mode = 0644,
1291 .proc_handler = proc_dointvec,
1292 .extra1 = &zero,
1293 },
1294 {
1295 .procname = "vfs_cache_pressure",
1296 .data = &sysctl_vfs_cache_pressure,
1297 .maxlen = sizeof(sysctl_vfs_cache_pressure),
1298 .mode = 0644,
1299 .proc_handler = proc_dointvec,
1300 .extra1 = &zero,
1301 },
1302 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1303 {
1304 .procname = "legacy_va_layout",
1305 .data = &sysctl_legacy_va_layout,
1306 .maxlen = sizeof(sysctl_legacy_va_layout),
1307 .mode = 0644,
1308 .proc_handler = proc_dointvec,
1309 .extra1 = &zero,
1310 },
1311 #endif
1312 #ifdef CONFIG_NUMA
1313 {
1314 .procname = "zone_reclaim_mode",
1315 .data = &zone_reclaim_mode,
1316 .maxlen = sizeof(zone_reclaim_mode),
1317 .mode = 0644,
1318 .proc_handler = proc_dointvec,
1319 .extra1 = &zero,
1320 },
1321 {
1322 .procname = "min_unmapped_ratio",
1323 .data = &sysctl_min_unmapped_ratio,
1324 .maxlen = sizeof(sysctl_min_unmapped_ratio),
1325 .mode = 0644,
1326 .proc_handler = sysctl_min_unmapped_ratio_sysctl_handler,
1327 .extra1 = &zero,
1328 .extra2 = &one_hundred,
1329 },
1330 {
1331 .procname = "min_slab_ratio",
1332 .data = &sysctl_min_slab_ratio,
1333 .maxlen = sizeof(sysctl_min_slab_ratio),
1334 .mode = 0644,
1335 .proc_handler = sysctl_min_slab_ratio_sysctl_handler,
1336 .extra1 = &zero,
1337 .extra2 = &one_hundred,
1338 },
1339 #endif
1340 #ifdef CONFIG_SMP
1341 {
1342 .procname = "stat_interval",
1343 .data = &sysctl_stat_interval,
1344 .maxlen = sizeof(sysctl_stat_interval),
1345 .mode = 0644,
1346 .proc_handler = proc_dointvec_jiffies,
1347 },
1348 #endif
1349 #ifdef CONFIG_MMU
1350 {
1351 .procname = "mmap_min_addr",
1352 .data = &dac_mmap_min_addr,
1353 .maxlen = sizeof(unsigned long),
1354 .mode = 0644,
1355 .proc_handler = mmap_min_addr_handler,
1356 },
1357 #endif
1358 #ifdef CONFIG_NUMA
1359 {
1360 .procname = "numa_zonelist_order",
1361 .data = &numa_zonelist_order,
1362 .maxlen = NUMA_ZONELIST_ORDER_LEN,
1363 .mode = 0644,
1364 .proc_handler = numa_zonelist_order_handler,
1365 },
1366 #endif
1367 #if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
1368 (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
1369 {
1370 .procname = "vdso_enabled",
1371 .data = &vdso_enabled,
1372 .maxlen = sizeof(vdso_enabled),
1373 .mode = 0644,
1374 .proc_handler = proc_dointvec,
1375 .extra1 = &zero,
1376 },
1377 #endif
1378 #ifdef CONFIG_HIGHMEM
1379 {
1380 .procname = "highmem_is_dirtyable",
1381 .data = &vm_highmem_is_dirtyable,
1382 .maxlen = sizeof(vm_highmem_is_dirtyable),
1383 .mode = 0644,
1384 .proc_handler = proc_dointvec_minmax,
1385 .extra1 = &zero,
1386 .extra2 = &one,
1387 },
1388 #endif
1389 {
1390 .procname = "scan_unevictable_pages",
1391 .data = &scan_unevictable_pages,
1392 .maxlen = sizeof(scan_unevictable_pages),
1393 .mode = 0644,
1394 .proc_handler = scan_unevictable_handler,
1395 },
1396 #ifdef CONFIG_MEMORY_FAILURE
1397 {
1398 .procname = "memory_failure_early_kill",
1399 .data = &sysctl_memory_failure_early_kill,
1400 .maxlen = sizeof(sysctl_memory_failure_early_kill),
1401 .mode = 0644,
1402 .proc_handler = proc_dointvec_minmax,
1403 .extra1 = &zero,
1404 .extra2 = &one,
1405 },
1406 {
1407 .procname = "memory_failure_recovery",
1408 .data = &sysctl_memory_failure_recovery,
1409 .maxlen = sizeof(sysctl_memory_failure_recovery),
1410 .mode = 0644,
1411 .proc_handler = proc_dointvec_minmax,
1412 .extra1 = &zero,
1413 .extra2 = &one,
1414 },
1415 #endif
1416 { }
1417 };
1418
1419 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1420 static struct ctl_table binfmt_misc_table[] = {
1421 { }
1422 };
1423 #endif
1424
1425 static struct ctl_table fs_table[] = {
1426 {
1427 .procname = "inode-nr",
1428 .data = &inodes_stat,
1429 .maxlen = 2*sizeof(int),
1430 .mode = 0444,
1431 .proc_handler = proc_nr_inodes,
1432 },
1433 {
1434 .procname = "inode-state",
1435 .data = &inodes_stat,
1436 .maxlen = 7*sizeof(int),
1437 .mode = 0444,
1438 .proc_handler = proc_nr_inodes,
1439 },
1440 {
1441 .procname = "file-nr",
1442 .data = &files_stat,
1443 .maxlen = sizeof(files_stat),
1444 .mode = 0444,
1445 .proc_handler = proc_nr_files,
1446 },
1447 {
1448 .procname = "file-max",
1449 .data = &files_stat.max_files,
1450 .maxlen = sizeof(files_stat.max_files),
1451 .mode = 0644,
1452 .proc_handler = proc_doulongvec_minmax,
1453 },
1454 {
1455 .procname = "nr_open",
1456 .data = &sysctl_nr_open,
1457 .maxlen = sizeof(int),
1458 .mode = 0644,
1459 .proc_handler = proc_dointvec_minmax,
1460 .extra1 = &sysctl_nr_open_min,
1461 .extra2 = &sysctl_nr_open_max,
1462 },
1463 {
1464 .procname = "dentry-state",
1465 .data = &dentry_stat,
1466 .maxlen = 6*sizeof(int),
1467 .mode = 0444,
1468 .proc_handler = proc_nr_dentry,
1469 },
1470 {
1471 .procname = "overflowuid",
1472 .data = &fs_overflowuid,
1473 .maxlen = sizeof(int),
1474 .mode = 0644,
1475 .proc_handler = proc_dointvec_minmax,
1476 .extra1 = &minolduid,
1477 .extra2 = &maxolduid,
1478 },
1479 {
1480 .procname = "overflowgid",
1481 .data = &fs_overflowgid,
1482 .maxlen = sizeof(int),
1483 .mode = 0644,
1484 .proc_handler = proc_dointvec_minmax,
1485 .extra1 = &minolduid,
1486 .extra2 = &maxolduid,
1487 },
1488 #ifdef CONFIG_FILE_LOCKING
1489 {
1490 .procname = "leases-enable",
1491 .data = &leases_enable,
1492 .maxlen = sizeof(int),
1493 .mode = 0644,
1494 .proc_handler = proc_dointvec,
1495 },
1496 #endif
1497 #ifdef CONFIG_DNOTIFY
1498 {
1499 .procname = "dir-notify-enable",
1500 .data = &dir_notify_enable,
1501 .maxlen = sizeof(int),
1502 .mode = 0644,
1503 .proc_handler = proc_dointvec,
1504 },
1505 #endif
1506 #ifdef CONFIG_MMU
1507 #ifdef CONFIG_FILE_LOCKING
1508 {
1509 .procname = "lease-break-time",
1510 .data = &lease_break_time,
1511 .maxlen = sizeof(int),
1512 .mode = 0644,
1513 .proc_handler = proc_dointvec,
1514 },
1515 #endif
1516 #ifdef CONFIG_AIO
1517 {
1518 .procname = "aio-nr",
1519 .data = &aio_nr,
1520 .maxlen = sizeof(aio_nr),
1521 .mode = 0444,
1522 .proc_handler = proc_doulongvec_minmax,
1523 },
1524 {
1525 .procname = "aio-max-nr",
1526 .data = &aio_max_nr,
1527 .maxlen = sizeof(aio_max_nr),
1528 .mode = 0644,
1529 .proc_handler = proc_doulongvec_minmax,
1530 },
1531 #endif /* CONFIG_AIO */
1532 #ifdef CONFIG_INOTIFY_USER
1533 {
1534 .procname = "inotify",
1535 .mode = 0555,
1536 .child = inotify_table,
1537 },
1538 #endif
1539 #ifdef CONFIG_EPOLL
1540 {
1541 .procname = "epoll",
1542 .mode = 0555,
1543 .child = epoll_table,
1544 },
1545 #endif
1546 #endif
1547 {
1548 .procname = "protected_symlinks",
1549 .data = &sysctl_protected_symlinks,
1550 .maxlen = sizeof(int),
1551 .mode = 0600,
1552 .proc_handler = proc_dointvec_minmax,
1553 .extra1 = &zero,
1554 .extra2 = &one,
1555 },
1556 {
1557 .procname = "protected_hardlinks",
1558 .data = &sysctl_protected_hardlinks,
1559 .maxlen = sizeof(int),
1560 .mode = 0600,
1561 .proc_handler = proc_dointvec_minmax,
1562 .extra1 = &zero,
1563 .extra2 = &one,
1564 },
1565 {
1566 .procname = "suid_dumpable",
1567 .data = &suid_dumpable,
1568 .maxlen = sizeof(int),
1569 .mode = 0644,
1570 .proc_handler = proc_dointvec_minmax_coredump,
1571 .extra1 = &zero,
1572 .extra2 = &two,
1573 },
1574 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1575 {
1576 .procname = "binfmt_misc",
1577 .mode = 0555,
1578 .child = binfmt_misc_table,
1579 },
1580 #endif
1581 {
1582 .procname = "pipe-max-size",
1583 .data = &pipe_max_size,
1584 .maxlen = sizeof(int),
1585 .mode = 0644,
1586 .proc_handler = &pipe_proc_fn,
1587 .extra1 = &pipe_min_size,
1588 },
1589 { }
1590 };
1591
1592 static struct ctl_table debug_table[] = {
1593 #ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
1594 {
1595 .procname = "exception-trace",
1596 .data = &show_unhandled_signals,
1597 .maxlen = sizeof(int),
1598 .mode = 0644,
1599 .proc_handler = proc_dointvec
1600 },
1601 #endif
1602 #if defined(CONFIG_OPTPROBES)
1603 {
1604 .procname = "kprobes-optimization",
1605 .data = &sysctl_kprobes_optimization,
1606 .maxlen = sizeof(int),
1607 .mode = 0644,
1608 .proc_handler = proc_kprobes_optimization_handler,
1609 .extra1 = &zero,
1610 .extra2 = &one,
1611 },
1612 #endif
1613 { }
1614 };
1615
1616 static struct ctl_table dev_table[] = {
1617 { }
1618 };
1619
1620 int __init sysctl_init(void)
1621 {
1622 struct ctl_table_header *hdr;
1623
1624 hdr = register_sysctl_table(sysctl_base_table);
1625 kmemleak_not_leak(hdr);
1626 return 0;
1627 }
1628
1629 #endif /* CONFIG_SYSCTL */
1630
1631 /*
1632 * /proc/sys support
1633 */
1634
1635 #ifdef CONFIG_PROC_SYSCTL
1636
1637 static int _proc_do_string(void* data, int maxlen, int write,
1638 void __user *buffer,
1639 size_t *lenp, loff_t *ppos)
1640 {
1641 size_t len;
1642 char __user *p;
1643 char c;
1644
1645 if (!data || !maxlen || !*lenp) {
1646 *lenp = 0;
1647 return 0;
1648 }
1649
1650 if (write) {
1651 len = 0;
1652 p = buffer;
1653 while (len < *lenp) {
1654 if (get_user(c, p++))
1655 return -EFAULT;
1656 if (c == 0 || c == '\n')
1657 break;
1658 len++;
1659 }
1660 if (len >= maxlen)
1661 len = maxlen-1;
1662 if(copy_from_user(data, buffer, len))
1663 return -EFAULT;
1664 ((char *) data)[len] = 0;
1665 *ppos += *lenp;
1666 } else {
1667 len = strlen(data);
1668 if (len > maxlen)
1669 len = maxlen;
1670
1671 if (*ppos > len) {
1672 *lenp = 0;
1673 return 0;
1674 }
1675
1676 data += *ppos;
1677 len -= *ppos;
1678
1679 if (len > *lenp)
1680 len = *lenp;
1681 if (len)
1682 if(copy_to_user(buffer, data, len))
1683 return -EFAULT;
1684 if (len < *lenp) {
1685 if(put_user('\n', ((char __user *) buffer) + len))
1686 return -EFAULT;
1687 len++;
1688 }
1689 *lenp = len;
1690 *ppos += len;
1691 }
1692 return 0;
1693 }
1694
1695 /**
1696 * proc_dostring - read a string sysctl
1697 * @table: the sysctl table
1698 * @write: %TRUE if this is a write to the sysctl file
1699 * @buffer: the user buffer
1700 * @lenp: the size of the user buffer
1701 * @ppos: file position
1702 *
1703 * Reads/writes a string from/to the user buffer. If the kernel
1704 * buffer provided is not large enough to hold the string, the
1705 * string is truncated. The copied string is %NULL-terminated.
1706 * If the string is being read by the user process, it is copied
1707 * and a newline '\n' is added. It is truncated if the buffer is
1708 * not large enough.
1709 *
1710 * Returns 0 on success.
1711 */
1712 int proc_dostring(struct ctl_table *table, int write,
1713 void __user *buffer, size_t *lenp, loff_t *ppos)
1714 {
1715 return _proc_do_string(table->data, table->maxlen, write,
1716 buffer, lenp, ppos);
1717 }
1718
1719 static size_t proc_skip_spaces(char **buf)
1720 {
1721 size_t ret;
1722 char *tmp = skip_spaces(*buf);
1723 ret = tmp - *buf;
1724 *buf = tmp;
1725 return ret;
1726 }
1727
1728 static void proc_skip_char(char **buf, size_t *size, const char v)
1729 {
1730 while (*size) {
1731 if (**buf != v)
1732 break;
1733 (*size)--;
1734 (*buf)++;
1735 }
1736 }
1737
1738 #define TMPBUFLEN 22
1739 /**
1740 * proc_get_long - reads an ASCII formatted integer from a user buffer
1741 *
1742 * @buf: a kernel buffer
1743 * @size: size of the kernel buffer
1744 * @val: this is where the number will be stored
1745 * @neg: set to %TRUE if number is negative
1746 * @perm_tr: a vector which contains the allowed trailers
1747 * @perm_tr_len: size of the perm_tr vector
1748 * @tr: pointer to store the trailer character
1749 *
1750 * In case of success %0 is returned and @buf and @size are updated with
1751 * the amount of bytes read. If @tr is non-NULL and a trailing
1752 * character exists (size is non-zero after returning from this
1753 * function), @tr is updated with the trailing character.
1754 */
1755 static int proc_get_long(char **buf, size_t *size,
1756 unsigned long *val, bool *neg,
1757 const char *perm_tr, unsigned perm_tr_len, char *tr)
1758 {
1759 int len;
1760 char *p, tmp[TMPBUFLEN];
1761
1762 if (!*size)
1763 return -EINVAL;
1764
1765 len = *size;
1766 if (len > TMPBUFLEN - 1)
1767 len = TMPBUFLEN - 1;
1768
1769 memcpy(tmp, *buf, len);
1770
1771 tmp[len] = 0;
1772 p = tmp;
1773 if (*p == '-' && *size > 1) {
1774 *neg = true;
1775 p++;
1776 } else
1777 *neg = false;
1778 if (!isdigit(*p))
1779 return -EINVAL;
1780
1781 *val = simple_strtoul(p, &p, 0);
1782
1783 len = p - tmp;
1784
1785 /* We don't know if the next char is whitespace thus we may accept
1786 * invalid integers (e.g. 1234...a) or two integers instead of one
1787 * (e.g. 123...1). So lets not allow such large numbers. */
1788 if (len == TMPBUFLEN - 1)
1789 return -EINVAL;
1790
1791 if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
1792 return -EINVAL;
1793
1794 if (tr && (len < *size))
1795 *tr = *p;
1796
1797 *buf += len;
1798 *size -= len;
1799
1800 return 0;
1801 }
1802
1803 /**
1804 * proc_put_long - converts an integer to a decimal ASCII formatted string
1805 *
1806 * @buf: the user buffer
1807 * @size: the size of the user buffer
1808 * @val: the integer to be converted
1809 * @neg: sign of the number, %TRUE for negative
1810 *
1811 * In case of success %0 is returned and @buf and @size are updated with
1812 * the amount of bytes written.
1813 */
1814 static int proc_put_long(void __user **buf, size_t *size, unsigned long val,
1815 bool neg)
1816 {
1817 int len;
1818 char tmp[TMPBUFLEN], *p = tmp;
1819
1820 sprintf(p, "%s%lu", neg ? "-" : "", val);
1821 len = strlen(tmp);
1822 if (len > *size)
1823 len = *size;
1824 if (copy_to_user(*buf, tmp, len))
1825 return -EFAULT;
1826 *size -= len;
1827 *buf += len;
1828 return 0;
1829 }
1830 #undef TMPBUFLEN
1831
1832 static int proc_put_char(void __user **buf, size_t *size, char c)
1833 {
1834 if (*size) {
1835 char __user **buffer = (char __user **)buf;
1836 if (put_user(c, *buffer))
1837 return -EFAULT;
1838 (*size)--, (*buffer)++;
1839 *buf = *buffer;
1840 }
1841 return 0;
1842 }
1843
1844 static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
1845 int *valp,
1846 int write, void *data)
1847 {
1848 if (write) {
1849 *valp = *negp ? -*lvalp : *lvalp;
1850 } else {
1851 int val = *valp;
1852 if (val < 0) {
1853 *negp = true;
1854 *lvalp = (unsigned long)-val;
1855 } else {
1856 *negp = false;
1857 *lvalp = (unsigned long)val;
1858 }
1859 }
1860 return 0;
1861 }
1862
1863 static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
1864
1865 static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
1866 int write, void __user *buffer,
1867 size_t *lenp, loff_t *ppos,
1868 int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
1869 int write, void *data),
1870 void *data)
1871 {
1872 int *i, vleft, first = 1, err = 0;
1873 unsigned long page = 0;
1874 size_t left;
1875 char *kbuf;
1876
1877 if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
1878 *lenp = 0;
1879 return 0;
1880 }
1881
1882 i = (int *) tbl_data;
1883 vleft = table->maxlen / sizeof(*i);
1884 left = *lenp;
1885
1886 if (!conv)
1887 conv = do_proc_dointvec_conv;
1888
1889 if (write) {
1890 if (left > PAGE_SIZE - 1)
1891 left = PAGE_SIZE - 1;
1892 page = __get_free_page(GFP_TEMPORARY);
1893 kbuf = (char *) page;
1894 if (!kbuf)
1895 return -ENOMEM;
1896 if (copy_from_user(kbuf, buffer, left)) {
1897 err = -EFAULT;
1898 goto free;
1899 }
1900 kbuf[left] = 0;
1901 }
1902
1903 for (; left && vleft--; i++, first=0) {
1904 unsigned long lval;
1905 bool neg;
1906
1907 if (write) {
1908 left -= proc_skip_spaces(&kbuf);
1909
1910 if (!left)
1911 break;
1912 err = proc_get_long(&kbuf, &left, &lval, &neg,
1913 proc_wspace_sep,
1914 sizeof(proc_wspace_sep), NULL);
1915 if (err)
1916 break;
1917 if (conv(&neg, &lval, i, 1, data)) {
1918 err = -EINVAL;
1919 break;
1920 }
1921 } else {
1922 if (conv(&neg, &lval, i, 0, data)) {
1923 err = -EINVAL;
1924 break;
1925 }
1926 if (!first)
1927 err = proc_put_char(&buffer, &left, '\t');
1928 if (err)
1929 break;
1930 err = proc_put_long(&buffer, &left, lval, neg);
1931 if (err)
1932 break;
1933 }
1934 }
1935
1936 if (!write && !first && left && !err)
1937 err = proc_put_char(&buffer, &left, '\n');
1938 if (write && !err && left)
1939 left -= proc_skip_spaces(&kbuf);
1940 free:
1941 if (write) {
1942 free_page(page);
1943 if (first)
1944 return err ? : -EINVAL;
1945 }
1946 *lenp -= left;
1947 *ppos += *lenp;
1948 return err;
1949 }
1950
1951 static int do_proc_dointvec(struct ctl_table *table, int write,
1952 void __user *buffer, size_t *lenp, loff_t *ppos,
1953 int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
1954 int write, void *data),
1955 void *data)
1956 {
1957 return __do_proc_dointvec(table->data, table, write,
1958 buffer, lenp, ppos, conv, data);
1959 }
1960
1961 /**
1962 * proc_dointvec - read a vector of integers
1963 * @table: the sysctl table
1964 * @write: %TRUE if this is a write to the sysctl file
1965 * @buffer: the user buffer
1966 * @lenp: the size of the user buffer
1967 * @ppos: file position
1968 *
1969 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1970 * values from/to the user buffer, treated as an ASCII string.
1971 *
1972 * Returns 0 on success.
1973 */
1974 int proc_dointvec(struct ctl_table *table, int write,
1975 void __user *buffer, size_t *lenp, loff_t *ppos)
1976 {
1977 return do_proc_dointvec(table,write,buffer,lenp,ppos,
1978 NULL,NULL);
1979 }
1980
1981 /*
1982 * Taint values can only be increased
1983 * This means we can safely use a temporary.
1984 */
1985 static int proc_taint(struct ctl_table *table, int write,
1986 void __user *buffer, size_t *lenp, loff_t *ppos)
1987 {
1988 struct ctl_table t;
1989 unsigned long tmptaint = get_taint();
1990 int err;
1991
1992 if (write && !capable(CAP_SYS_ADMIN))
1993 return -EPERM;
1994
1995 t = *table;
1996 t.data = &tmptaint;
1997 err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
1998 if (err < 0)
1999 return err;
2000
2001 if (write) {
2002 /*
2003 * Poor man's atomic or. Not worth adding a primitive
2004 * to everyone's atomic.h for this
2005 */
2006 int i;
2007 for (i = 0; i < BITS_PER_LONG && tmptaint >> i; i++) {
2008 if ((tmptaint >> i) & 1)
2009 add_taint(i);
2010 }
2011 }
2012
2013 return err;
2014 }
2015
2016 #ifdef CONFIG_PRINTK
2017 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
2018 void __user *buffer, size_t *lenp, loff_t *ppos)
2019 {
2020 if (write && !capable(CAP_SYS_ADMIN))
2021 return -EPERM;
2022
2023 return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2024 }
2025 #endif
2026
2027 struct do_proc_dointvec_minmax_conv_param {
2028 int *min;
2029 int *max;
2030 };
2031
2032 static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
2033 int *valp,
2034 int write, void *data)
2035 {
2036 struct do_proc_dointvec_minmax_conv_param *param = data;
2037 if (write) {
2038 int val = *negp ? -*lvalp : *lvalp;
2039 if ((param->min && *param->min > val) ||
2040 (param->max && *param->max < val))
2041 return -EINVAL;
2042 *valp = val;
2043 } else {
2044 int val = *valp;
2045 if (val < 0) {
2046 *negp = true;
2047 *lvalp = (unsigned long)-val;
2048 } else {
2049 *negp = false;
2050 *lvalp = (unsigned long)val;
2051 }
2052 }
2053 return 0;
2054 }
2055
2056 /**
2057 * proc_dointvec_minmax - read a vector of integers with min/max values
2058 * @table: the sysctl table
2059 * @write: %TRUE if this is a write to the sysctl file
2060 * @buffer: the user buffer
2061 * @lenp: the size of the user buffer
2062 * @ppos: file position
2063 *
2064 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2065 * values from/to the user buffer, treated as an ASCII string.
2066 *
2067 * This routine will ensure the values are within the range specified by
2068 * table->extra1 (min) and table->extra2 (max).
2069 *
2070 * Returns 0 on success.
2071 */
2072 int proc_dointvec_minmax(struct ctl_table *table, int write,
2073 void __user *buffer, size_t *lenp, loff_t *ppos)
2074 {
2075 struct do_proc_dointvec_minmax_conv_param param = {
2076 .min = (int *) table->extra1,
2077 .max = (int *) table->extra2,
2078 };
2079 return do_proc_dointvec(table, write, buffer, lenp, ppos,
2080 do_proc_dointvec_minmax_conv, &param);
2081 }
2082
2083 static void validate_coredump_safety(void)
2084 {
2085 #ifdef CONFIG_COREDUMP
2086 if (suid_dumpable == SUID_DUMPABLE_SAFE &&
2087 core_pattern[0] != '/' && core_pattern[0] != '|') {
2088 printk(KERN_WARNING "Unsafe core_pattern used with "\
2089 "suid_dumpable=2. Pipe handler or fully qualified "\
2090 "core dump path required.\n");
2091 }
2092 #endif
2093 }
2094
2095 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
2096 void __user *buffer, size_t *lenp, loff_t *ppos)
2097 {
2098 int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2099 if (!error)
2100 validate_coredump_safety();
2101 return error;
2102 }
2103
2104 #ifdef CONFIG_COREDUMP
2105 static int proc_dostring_coredump(struct ctl_table *table, int write,
2106 void __user *buffer, size_t *lenp, loff_t *ppos)
2107 {
2108 int error = proc_dostring(table, write, buffer, lenp, ppos);
2109 if (!error)
2110 validate_coredump_safety();
2111 return error;
2112 }
2113 #endif
2114
2115 static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table, int write,
2116 void __user *buffer,
2117 size_t *lenp, loff_t *ppos,
2118 unsigned long convmul,
2119 unsigned long convdiv)
2120 {
2121 unsigned long *i, *min, *max;
2122 int vleft, first = 1, err = 0;
2123 unsigned long page = 0;
2124 size_t left;
2125 char *kbuf;
2126
2127 if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
2128 *lenp = 0;
2129 return 0;
2130 }
2131
2132 i = (unsigned long *) data;
2133 min = (unsigned long *) table->extra1;
2134 max = (unsigned long *) table->extra2;
2135 vleft = table->maxlen / sizeof(unsigned long);
2136 left = *lenp;
2137
2138 if (write) {
2139 if (left > PAGE_SIZE - 1)
2140 left = PAGE_SIZE - 1;
2141 page = __get_free_page(GFP_TEMPORARY);
2142 kbuf = (char *) page;
2143 if (!kbuf)
2144 return -ENOMEM;
2145 if (copy_from_user(kbuf, buffer, left)) {
2146 err = -EFAULT;
2147 goto free;
2148 }
2149 kbuf[left] = 0;
2150 }
2151
2152 for (; left && vleft--; i++, first = 0) {
2153 unsigned long val;
2154
2155 if (write) {
2156 bool neg;
2157
2158 left -= proc_skip_spaces(&kbuf);
2159
2160 err = proc_get_long(&kbuf, &left, &val, &neg,
2161 proc_wspace_sep,
2162 sizeof(proc_wspace_sep), NULL);
2163 if (err)
2164 break;
2165 if (neg)
2166 continue;
2167 if ((min && val < *min) || (max && val > *max))
2168 continue;
2169 *i = val;
2170 } else {
2171 val = convdiv * (*i) / convmul;
2172 if (!first)
2173 err = proc_put_char(&buffer, &left, '\t');
2174 err = proc_put_long(&buffer, &left, val, false);
2175 if (err)
2176 break;
2177 }
2178 }
2179
2180 if (!write && !first && left && !err)
2181 err = proc_put_char(&buffer, &left, '\n');
2182 if (write && !err)
2183 left -= proc_skip_spaces(&kbuf);
2184 free:
2185 if (write) {
2186 free_page(page);
2187 if (first)
2188 return err ? : -EINVAL;
2189 }
2190 *lenp -= left;
2191 *ppos += *lenp;
2192 return err;
2193 }
2194
2195 static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
2196 void __user *buffer,
2197 size_t *lenp, loff_t *ppos,
2198 unsigned long convmul,
2199 unsigned long convdiv)
2200 {
2201 return __do_proc_doulongvec_minmax(table->data, table, write,
2202 buffer, lenp, ppos, convmul, convdiv);
2203 }
2204
2205 /**
2206 * proc_doulongvec_minmax - read a vector of long integers with min/max values
2207 * @table: the sysctl table
2208 * @write: %TRUE if this is a write to the sysctl file
2209 * @buffer: the user buffer
2210 * @lenp: the size of the user buffer
2211 * @ppos: file position
2212 *
2213 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2214 * values from/to the user buffer, treated as an ASCII string.
2215 *
2216 * This routine will ensure the values are within the range specified by
2217 * table->extra1 (min) and table->extra2 (max).
2218 *
2219 * Returns 0 on success.
2220 */
2221 int proc_doulongvec_minmax(struct ctl_table *table, int write,
2222 void __user *buffer, size_t *lenp, loff_t *ppos)
2223 {
2224 return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
2225 }
2226
2227 /**
2228 * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
2229 * @table: the sysctl table
2230 * @write: %TRUE if this is a write to the sysctl file
2231 * @buffer: the user buffer
2232 * @lenp: the size of the user buffer
2233 * @ppos: file position
2234 *
2235 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2236 * values from/to the user buffer, treated as an ASCII string. The values
2237 * are treated as milliseconds, and converted to jiffies when they are stored.
2238 *
2239 * This routine will ensure the values are within the range specified by
2240 * table->extra1 (min) and table->extra2 (max).
2241 *
2242 * Returns 0 on success.
2243 */
2244 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
2245 void __user *buffer,
2246 size_t *lenp, loff_t *ppos)
2247 {
2248 return do_proc_doulongvec_minmax(table, write, buffer,
2249 lenp, ppos, HZ, 1000l);
2250 }
2251
2252
2253 static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
2254 int *valp,
2255 int write, void *data)
2256 {
2257 if (write) {
2258 if (*lvalp > LONG_MAX / HZ)
2259 return 1;
2260 *valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
2261 } else {
2262 int val = *valp;
2263 unsigned long lval;
2264 if (val < 0) {
2265 *negp = true;
2266 lval = (unsigned long)-val;
2267 } else {
2268 *negp = false;
2269 lval = (unsigned long)val;
2270 }
2271 *lvalp = lval / HZ;
2272 }
2273 return 0;
2274 }
2275
2276 static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
2277 int *valp,
2278 int write, void *data)
2279 {
2280 if (write) {
2281 if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
2282 return 1;
2283 *valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
2284 } else {
2285 int val = *valp;
2286 unsigned long lval;
2287 if (val < 0) {
2288 *negp = true;
2289 lval = (unsigned long)-val;
2290 } else {
2291 *negp = false;
2292 lval = (unsigned long)val;
2293 }
2294 *lvalp = jiffies_to_clock_t(lval);
2295 }
2296 return 0;
2297 }
2298
2299 static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
2300 int *valp,
2301 int write, void *data)
2302 {
2303 if (write) {
2304 *valp = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
2305 } else {
2306 int val = *valp;
2307 unsigned long lval;
2308 if (val < 0) {
2309 *negp = true;
2310 lval = (unsigned long)-val;
2311 } else {
2312 *negp = false;
2313 lval = (unsigned long)val;
2314 }
2315 *lvalp = jiffies_to_msecs(lval);
2316 }
2317 return 0;
2318 }
2319
2320 /**
2321 * proc_dointvec_jiffies - read a vector of integers as seconds
2322 * @table: the sysctl table
2323 * @write: %TRUE if this is a write to the sysctl file
2324 * @buffer: the user buffer
2325 * @lenp: the size of the user buffer
2326 * @ppos: file position
2327 *
2328 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2329 * values from/to the user buffer, treated as an ASCII string.
2330 * The values read are assumed to be in seconds, and are converted into
2331 * jiffies.
2332 *
2333 * Returns 0 on success.
2334 */
2335 int proc_dointvec_jiffies(struct ctl_table *table, int write,
2336 void __user *buffer, size_t *lenp, loff_t *ppos)
2337 {
2338 return do_proc_dointvec(table,write,buffer,lenp,ppos,
2339 do_proc_dointvec_jiffies_conv,NULL);
2340 }
2341
2342 /**
2343 * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
2344 * @table: the sysctl table
2345 * @write: %TRUE if this is a write to the sysctl file
2346 * @buffer: the user buffer
2347 * @lenp: the size of the user buffer
2348 * @ppos: pointer to the file position
2349 *
2350 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2351 * values from/to the user buffer, treated as an ASCII string.
2352 * The values read are assumed to be in 1/USER_HZ seconds, and
2353 * are converted into jiffies.
2354 *
2355 * Returns 0 on success.
2356 */
2357 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
2358 void __user *buffer, size_t *lenp, loff_t *ppos)
2359 {
2360 return do_proc_dointvec(table,write,buffer,lenp,ppos,
2361 do_proc_dointvec_userhz_jiffies_conv,NULL);
2362 }
2363
2364 /**
2365 * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
2366 * @table: the sysctl table
2367 * @write: %TRUE if this is a write to the sysctl file
2368 * @buffer: the user buffer
2369 * @lenp: the size of the user buffer
2370 * @ppos: file position
2371 * @ppos: the current position in the file
2372 *
2373 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2374 * values from/to the user buffer, treated as an ASCII string.
2375 * The values read are assumed to be in 1/1000 seconds, and
2376 * are converted into jiffies.
2377 *
2378 * Returns 0 on success.
2379 */
2380 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
2381 void __user *buffer, size_t *lenp, loff_t *ppos)
2382 {
2383 return do_proc_dointvec(table, write, buffer, lenp, ppos,
2384 do_proc_dointvec_ms_jiffies_conv, NULL);
2385 }
2386
2387 static int proc_do_cad_pid(struct ctl_table *table, int write,
2388 void __user *buffer, size_t *lenp, loff_t *ppos)
2389 {
2390 struct pid *new_pid;
2391 pid_t tmp;
2392 int r;
2393
2394 tmp = pid_vnr(cad_pid);
2395
2396 r = __do_proc_dointvec(&tmp, table, write, buffer,
2397 lenp, ppos, NULL, NULL);
2398 if (r || !write)
2399 return r;
2400
2401 new_pid = find_get_pid(tmp);
2402 if (!new_pid)
2403 return -ESRCH;
2404
2405 put_pid(xchg(&cad_pid, new_pid));
2406 return 0;
2407 }
2408
2409 /**
2410 * proc_do_large_bitmap - read/write from/to a large bitmap
2411 * @table: the sysctl table
2412 * @write: %TRUE if this is a write to the sysctl file
2413 * @buffer: the user buffer
2414 * @lenp: the size of the user buffer
2415 * @ppos: file position
2416 *
2417 * The bitmap is stored at table->data and the bitmap length (in bits)
2418 * in table->maxlen.
2419 *
2420 * We use a range comma separated format (e.g. 1,3-4,10-10) so that
2421 * large bitmaps may be represented in a compact manner. Writing into
2422 * the file will clear the bitmap then update it with the given input.
2423 *
2424 * Returns 0 on success.
2425 */
2426 int proc_do_large_bitmap(struct ctl_table *table, int write,
2427 void __user *buffer, size_t *lenp, loff_t *ppos)
2428 {
2429 int err = 0;
2430 bool first = 1;
2431 size_t left = *lenp;
2432 unsigned long bitmap_len = table->maxlen;
2433 unsigned long *bitmap = (unsigned long *) table->data;
2434 unsigned long *tmp_bitmap = NULL;
2435 char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
2436
2437 if (!bitmap_len || !left || (*ppos && !write)) {
2438 *lenp = 0;
2439 return 0;
2440 }
2441
2442 if (write) {
2443 unsigned long page = 0;
2444 char *kbuf;
2445
2446 if (left > PAGE_SIZE - 1)
2447 left = PAGE_SIZE - 1;
2448
2449 page = __get_free_page(GFP_TEMPORARY);
2450 kbuf = (char *) page;
2451 if (!kbuf)
2452 return -ENOMEM;
2453 if (copy_from_user(kbuf, buffer, left)) {
2454 free_page(page);
2455 return -EFAULT;
2456 }
2457 kbuf[left] = 0;
2458
2459 tmp_bitmap = kzalloc(BITS_TO_LONGS(bitmap_len) * sizeof(unsigned long),
2460 GFP_KERNEL);
2461 if (!tmp_bitmap) {
2462 free_page(page);
2463 return -ENOMEM;
2464 }
2465 proc_skip_char(&kbuf, &left, '\n');
2466 while (!err && left) {
2467 unsigned long val_a, val_b;
2468 bool neg;
2469
2470 err = proc_get_long(&kbuf, &left, &val_a, &neg, tr_a,
2471 sizeof(tr_a), &c);
2472 if (err)
2473 break;
2474 if (val_a >= bitmap_len || neg) {
2475 err = -EINVAL;
2476 break;
2477 }
2478
2479 val_b = val_a;
2480 if (left) {
2481 kbuf++;
2482 left--;
2483 }
2484
2485 if (c == '-') {
2486 err = proc_get_long(&kbuf, &left, &val_b,
2487 &neg, tr_b, sizeof(tr_b),
2488 &c);
2489 if (err)
2490 break;
2491 if (val_b >= bitmap_len || neg ||
2492 val_a > val_b) {
2493 err = -EINVAL;
2494 break;
2495 }
2496 if (left) {
2497 kbuf++;
2498 left--;
2499 }
2500 }
2501
2502 bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
2503 first = 0;
2504 proc_skip_char(&kbuf, &left, '\n');
2505 }
2506 free_page(page);
2507 } else {
2508 unsigned long bit_a, bit_b = 0;
2509
2510 while (left) {
2511 bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
2512 if (bit_a >= bitmap_len)
2513 break;
2514 bit_b = find_next_zero_bit(bitmap, bitmap_len,
2515 bit_a + 1) - 1;
2516
2517 if (!first) {
2518 err = proc_put_char(&buffer, &left, ',');
2519 if (err)
2520 break;
2521 }
2522 err = proc_put_long(&buffer, &left, bit_a, false);
2523 if (err)
2524 break;
2525 if (bit_a != bit_b) {
2526 err = proc_put_char(&buffer, &left, '-');
2527 if (err)
2528 break;
2529 err = proc_put_long(&buffer, &left, bit_b, false);
2530 if (err)
2531 break;
2532 }
2533
2534 first = 0; bit_b++;
2535 }
2536 if (!err)
2537 err = proc_put_char(&buffer, &left, '\n');
2538 }
2539
2540 if (!err) {
2541 if (write) {
2542 if (*ppos)
2543 bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
2544 else
2545 bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
2546 }
2547 kfree(tmp_bitmap);
2548 *lenp -= left;
2549 *ppos += *lenp;
2550 return 0;
2551 } else {
2552 kfree(tmp_bitmap);
2553 return err;
2554 }
2555 }
2556
2557 #else /* CONFIG_PROC_SYSCTL */
2558
2559 int proc_dostring(struct ctl_table *table, int write,
2560 void __user *buffer, size_t *lenp, loff_t *ppos)
2561 {
2562 return -ENOSYS;
2563 }
2564
2565 int proc_dointvec(struct ctl_table *table, int write,
2566 void __user *buffer, size_t *lenp, loff_t *ppos)
2567 {
2568 return -ENOSYS;
2569 }
2570
2571 int proc_dointvec_minmax(struct ctl_table *table, int write,
2572 void __user *buffer, size_t *lenp, loff_t *ppos)
2573 {
2574 return -ENOSYS;
2575 }
2576
2577 int proc_dointvec_jiffies(struct ctl_table *table, int write,
2578 void __user *buffer, size_t *lenp, loff_t *ppos)
2579 {
2580 return -ENOSYS;
2581 }
2582
2583 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
2584 void __user *buffer, size_t *lenp, loff_t *ppos)
2585 {
2586 return -ENOSYS;
2587 }
2588
2589 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
2590 void __user *buffer, size_t *lenp, loff_t *ppos)
2591 {
2592 return -ENOSYS;
2593 }
2594
2595 int proc_doulongvec_minmax(struct ctl_table *table, int write,
2596 void __user *buffer, size_t *lenp, loff_t *ppos)
2597 {
2598 return -ENOSYS;
2599 }
2600
2601 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
2602 void __user *buffer,
2603 size_t *lenp, loff_t *ppos)
2604 {
2605 return -ENOSYS;
2606 }
2607
2608
2609 #endif /* CONFIG_PROC_SYSCTL */
2610
2611 /*
2612 * No sense putting this after each symbol definition, twice,
2613 * exception granted :-)
2614 */
2615 EXPORT_SYMBOL(proc_dointvec);
2616 EXPORT_SYMBOL(proc_dointvec_jiffies);
2617 EXPORT_SYMBOL(proc_dointvec_minmax);
2618 EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
2619 EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
2620 EXPORT_SYMBOL(proc_dostring);
2621 EXPORT_SYMBOL(proc_doulongvec_minmax);
2622 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
Linus' kernel tree