xref: /openbmc/linux/kernel/sysctl.c (revision cff11abeca78aa782378401ca2800bd2194aa14e)
1 // SPDX-License-Identifier: GPL-2.0-only
2 /*
3  * sysctl.c: General linux system control interface
4  *
5  * Begun 24 March 1995, Stephen Tweedie
6  * Added /proc support, Dec 1995
7  * Added bdflush entry and intvec min/max checking, 2/23/96, Tom Dyas.
8  * Added hooks for /proc/sys/net (minor, minor patch), 96/4/1, Mike Shaver.
9  * Added kernel/java-{interpreter,appletviewer}, 96/5/10, Mike Shaver.
10  * Dynamic registration fixes, Stephen Tweedie.
11  * Added kswapd-interval, ctrl-alt-del, printk stuff, 1/8/97, Chris Horn.
12  * Made sysctl support optional via CONFIG_SYSCTL, 1/10/97, Chris
13  *  Horn.
14  * Added proc_doulongvec_ms_jiffies_minmax, 09/08/99, Carlos H. Bauer.
15  * Added proc_doulongvec_minmax, 09/08/99, Carlos H. Bauer.
16  * Changed linked lists to use list.h instead of lists.h, 02/24/00, Bill
17  *  Wendling.
18  * The list_for_each() macro wasn't appropriate for the sysctl loop.
19  *  Removed it and replaced it with older style, 03/23/00, Bill Wendling
20  */
21 
22 #include <linux/module.h>
23 #include <linux/aio.h>
24 #include <linux/mm.h>
25 #include <linux/swap.h>
26 #include <linux/slab.h>
27 #include <linux/sysctl.h>
28 #include <linux/bitmap.h>
29 #include <linux/signal.h>
30 #include <linux/printk.h>
31 #include <linux/proc_fs.h>
32 #include <linux/security.h>
33 #include <linux/ctype.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/sched/coredump.h>
67 #include <linux/kexec.h>
68 #include <linux/bpf.h>
69 #include <linux/mount.h>
70 #include <linux/userfaultfd_k.h>
71 #include <linux/coredump.h>
72 #include <linux/latencytop.h>
73 #include <linux/pid.h>
74 
75 #include "../lib/kstrtox.h"
76 
77 #include <linux/uaccess.h>
78 #include <asm/processor.h>
79 
80 #ifdef CONFIG_X86
81 #include <asm/nmi.h>
82 #include <asm/stacktrace.h>
83 #include <asm/io.h>
84 #endif
85 #ifdef CONFIG_SPARC
86 #include <asm/setup.h>
87 #endif
88 #ifdef CONFIG_BSD_PROCESS_ACCT
89 #include <linux/acct.h>
90 #endif
91 #ifdef CONFIG_RT_MUTEXES
92 #include <linux/rtmutex.h>
93 #endif
94 #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
95 #include <linux/lockdep.h>
96 #endif
97 #ifdef CONFIG_CHR_DEV_SG
98 #include <scsi/sg.h>
99 #endif
100 #ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
101 #include <linux/stackleak.h>
102 #endif
103 #ifdef CONFIG_LOCKUP_DETECTOR
104 #include <linux/nmi.h>
105 #endif
106 
107 #if defined(CONFIG_SYSCTL)
108 
109 /* Constants used for minimum and  maximum */
110 #ifdef CONFIG_LOCKUP_DETECTOR
111 static int sixty = 60;
112 #endif
113 
114 static int __maybe_unused neg_one = -1;
115 static int __maybe_unused two = 2;
116 static int __maybe_unused four = 4;
117 static unsigned long zero_ul;
118 static unsigned long one_ul = 1;
119 static unsigned long long_max = LONG_MAX;
120 static int one_hundred = 100;
121 static int two_hundred = 200;
122 static int one_thousand = 1000;
123 #ifdef CONFIG_PRINTK
124 static int ten_thousand = 10000;
125 #endif
126 #ifdef CONFIG_PERF_EVENTS
127 static int six_hundred_forty_kb = 640 * 1024;
128 #endif
129 
130 /* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
131 static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
132 
133 /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
134 static int maxolduid = 65535;
135 static int minolduid;
136 
137 static int ngroups_max = NGROUPS_MAX;
138 static const int cap_last_cap = CAP_LAST_CAP;
139 
140 /*
141  * This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
142  * and hung_task_check_interval_secs
143  */
144 #ifdef CONFIG_DETECT_HUNG_TASK
145 static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
146 #endif
147 
148 #ifdef CONFIG_INOTIFY_USER
149 #include <linux/inotify.h>
150 #endif
151 
152 #ifdef CONFIG_PROC_SYSCTL
153 
154 /**
155  * enum sysctl_writes_mode - supported sysctl write modes
156  *
157  * @SYSCTL_WRITES_LEGACY: each write syscall must fully contain the sysctl value
158  *	to be written, and multiple writes on the same sysctl file descriptor
159  *	will rewrite the sysctl value, regardless of file position. No warning
160  *	is issued when the initial position is not 0.
161  * @SYSCTL_WRITES_WARN: same as above but warn when the initial file position is
162  *	not 0.
163  * @SYSCTL_WRITES_STRICT: writes to numeric sysctl entries must always be at
164  *	file position 0 and the value must be fully contained in the buffer
165  *	sent to the write syscall. If dealing with strings respect the file
166  *	position, but restrict this to the max length of the buffer, anything
167  *	passed the max length will be ignored. Multiple writes will append
168  *	to the buffer.
169  *
170  * These write modes control how current file position affects the behavior of
171  * updating sysctl values through the proc interface on each write.
172  */
173 enum sysctl_writes_mode {
174 	SYSCTL_WRITES_LEGACY		= -1,
175 	SYSCTL_WRITES_WARN		= 0,
176 	SYSCTL_WRITES_STRICT		= 1,
177 };
178 
179 static enum sysctl_writes_mode sysctl_writes_strict = SYSCTL_WRITES_STRICT;
180 #endif /* CONFIG_PROC_SYSCTL */
181 
182 #if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
183     defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
184 int sysctl_legacy_va_layout;
185 #endif
186 
187 #ifdef CONFIG_SCHED_DEBUG
188 static int min_sched_granularity_ns = 100000;		/* 100 usecs */
189 static int max_sched_granularity_ns = NSEC_PER_SEC;	/* 1 second */
190 static int min_wakeup_granularity_ns;			/* 0 usecs */
191 static int max_wakeup_granularity_ns = NSEC_PER_SEC;	/* 1 second */
192 #ifdef CONFIG_SMP
193 static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
194 static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
195 #endif /* CONFIG_SMP */
196 #endif /* CONFIG_SCHED_DEBUG */
197 
198 #ifdef CONFIG_COMPACTION
199 static int min_extfrag_threshold;
200 static int max_extfrag_threshold = 1000;
201 #endif
202 
203 #endif /* CONFIG_SYSCTL */
204 
205 #if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_SYSCTL)
206 static int bpf_stats_handler(struct ctl_table *table, int write,
207 			     void __user *buffer, size_t *lenp,
208 			     loff_t *ppos)
209 {
210 	struct static_key *key = (struct static_key *)table->data;
211 	static int saved_val;
212 	int val, ret;
213 	struct ctl_table tmp = {
214 		.data   = &val,
215 		.maxlen = sizeof(val),
216 		.mode   = table->mode,
217 		.extra1 = SYSCTL_ZERO,
218 		.extra2 = SYSCTL_ONE,
219 	};
220 
221 	if (write && !capable(CAP_SYS_ADMIN))
222 		return -EPERM;
223 
224 	mutex_lock(&bpf_stats_enabled_mutex);
225 	val = saved_val;
226 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
227 	if (write && !ret && val != saved_val) {
228 		if (val)
229 			static_key_slow_inc(key);
230 		else
231 			static_key_slow_dec(key);
232 		saved_val = val;
233 	}
234 	mutex_unlock(&bpf_stats_enabled_mutex);
235 	return ret;
236 }
237 #endif
238 
239 /*
240  * /proc/sys support
241  */
242 
243 #ifdef CONFIG_PROC_SYSCTL
244 
245 static int _proc_do_string(char *data, int maxlen, int write,
246 		char *buffer, size_t *lenp, loff_t *ppos)
247 {
248 	size_t len;
249 	char c, *p;
250 
251 	if (!data || !maxlen || !*lenp) {
252 		*lenp = 0;
253 		return 0;
254 	}
255 
256 	if (write) {
257 		if (sysctl_writes_strict == SYSCTL_WRITES_STRICT) {
258 			/* Only continue writes not past the end of buffer. */
259 			len = strlen(data);
260 			if (len > maxlen - 1)
261 				len = maxlen - 1;
262 
263 			if (*ppos > len)
264 				return 0;
265 			len = *ppos;
266 		} else {
267 			/* Start writing from beginning of buffer. */
268 			len = 0;
269 		}
270 
271 		*ppos += *lenp;
272 		p = buffer;
273 		while ((p - buffer) < *lenp && len < maxlen - 1) {
274 			c = *(p++);
275 			if (c == 0 || c == '\n')
276 				break;
277 			data[len++] = c;
278 		}
279 		data[len] = 0;
280 	} else {
281 		len = strlen(data);
282 		if (len > maxlen)
283 			len = maxlen;
284 
285 		if (*ppos > len) {
286 			*lenp = 0;
287 			return 0;
288 		}
289 
290 		data += *ppos;
291 		len  -= *ppos;
292 
293 		if (len > *lenp)
294 			len = *lenp;
295 		if (len)
296 			memcpy(buffer, data, len);
297 		if (len < *lenp) {
298 			buffer[len] = '\n';
299 			len++;
300 		}
301 		*lenp = len;
302 		*ppos += len;
303 	}
304 	return 0;
305 }
306 
307 static void warn_sysctl_write(struct ctl_table *table)
308 {
309 	pr_warn_once("%s wrote to %s when file position was not 0!\n"
310 		"This will not be supported in the future. To silence this\n"
311 		"warning, set kernel.sysctl_writes_strict = -1\n",
312 		current->comm, table->procname);
313 }
314 
315 /**
316  * proc_first_pos_non_zero_ignore - check if first position is allowed
317  * @ppos: file position
318  * @table: the sysctl table
319  *
320  * Returns true if the first position is non-zero and the sysctl_writes_strict
321  * mode indicates this is not allowed for numeric input types. String proc
322  * handlers can ignore the return value.
323  */
324 static bool proc_first_pos_non_zero_ignore(loff_t *ppos,
325 					   struct ctl_table *table)
326 {
327 	if (!*ppos)
328 		return false;
329 
330 	switch (sysctl_writes_strict) {
331 	case SYSCTL_WRITES_STRICT:
332 		return true;
333 	case SYSCTL_WRITES_WARN:
334 		warn_sysctl_write(table);
335 		return false;
336 	default:
337 		return false;
338 	}
339 }
340 
341 /**
342  * proc_dostring - read a string sysctl
343  * @table: the sysctl table
344  * @write: %TRUE if this is a write to the sysctl file
345  * @buffer: the user buffer
346  * @lenp: the size of the user buffer
347  * @ppos: file position
348  *
349  * Reads/writes a string from/to the user buffer. If the kernel
350  * buffer provided is not large enough to hold the string, the
351  * string is truncated. The copied string is %NULL-terminated.
352  * If the string is being read by the user process, it is copied
353  * and a newline '\n' is added. It is truncated if the buffer is
354  * not large enough.
355  *
356  * Returns 0 on success.
357  */
358 int proc_dostring(struct ctl_table *table, int write,
359 		  void *buffer, size_t *lenp, loff_t *ppos)
360 {
361 	if (write)
362 		proc_first_pos_non_zero_ignore(ppos, table);
363 
364 	return _proc_do_string(table->data, table->maxlen, write, buffer, lenp,
365 			ppos);
366 }
367 
368 static size_t proc_skip_spaces(char **buf)
369 {
370 	size_t ret;
371 	char *tmp = skip_spaces(*buf);
372 	ret = tmp - *buf;
373 	*buf = tmp;
374 	return ret;
375 }
376 
377 static void proc_skip_char(char **buf, size_t *size, const char v)
378 {
379 	while (*size) {
380 		if (**buf != v)
381 			break;
382 		(*size)--;
383 		(*buf)++;
384 	}
385 }
386 
387 /**
388  * strtoul_lenient - parse an ASCII formatted integer from a buffer and only
389  *                   fail on overflow
390  *
391  * @cp: kernel buffer containing the string to parse
392  * @endp: pointer to store the trailing characters
393  * @base: the base to use
394  * @res: where the parsed integer will be stored
395  *
396  * In case of success 0 is returned and @res will contain the parsed integer,
397  * @endp will hold any trailing characters.
398  * This function will fail the parse on overflow. If there wasn't an overflow
399  * the function will defer the decision what characters count as invalid to the
400  * caller.
401  */
402 static int strtoul_lenient(const char *cp, char **endp, unsigned int base,
403 			   unsigned long *res)
404 {
405 	unsigned long long result;
406 	unsigned int rv;
407 
408 	cp = _parse_integer_fixup_radix(cp, &base);
409 	rv = _parse_integer(cp, base, &result);
410 	if ((rv & KSTRTOX_OVERFLOW) || (result != (unsigned long)result))
411 		return -ERANGE;
412 
413 	cp += rv;
414 
415 	if (endp)
416 		*endp = (char *)cp;
417 
418 	*res = (unsigned long)result;
419 	return 0;
420 }
421 
422 #define TMPBUFLEN 22
423 /**
424  * proc_get_long - reads an ASCII formatted integer from a user buffer
425  *
426  * @buf: a kernel buffer
427  * @size: size of the kernel buffer
428  * @val: this is where the number will be stored
429  * @neg: set to %TRUE if number is negative
430  * @perm_tr: a vector which contains the allowed trailers
431  * @perm_tr_len: size of the perm_tr vector
432  * @tr: pointer to store the trailer character
433  *
434  * In case of success %0 is returned and @buf and @size are updated with
435  * the amount of bytes read. If @tr is non-NULL and a trailing
436  * character exists (size is non-zero after returning from this
437  * function), @tr is updated with the trailing character.
438  */
439 static int proc_get_long(char **buf, size_t *size,
440 			  unsigned long *val, bool *neg,
441 			  const char *perm_tr, unsigned perm_tr_len, char *tr)
442 {
443 	int len;
444 	char *p, tmp[TMPBUFLEN];
445 
446 	if (!*size)
447 		return -EINVAL;
448 
449 	len = *size;
450 	if (len > TMPBUFLEN - 1)
451 		len = TMPBUFLEN - 1;
452 
453 	memcpy(tmp, *buf, len);
454 
455 	tmp[len] = 0;
456 	p = tmp;
457 	if (*p == '-' && *size > 1) {
458 		*neg = true;
459 		p++;
460 	} else
461 		*neg = false;
462 	if (!isdigit(*p))
463 		return -EINVAL;
464 
465 	if (strtoul_lenient(p, &p, 0, val))
466 		return -EINVAL;
467 
468 	len = p - tmp;
469 
470 	/* We don't know if the next char is whitespace thus we may accept
471 	 * invalid integers (e.g. 1234...a) or two integers instead of one
472 	 * (e.g. 123...1). So lets not allow such large numbers. */
473 	if (len == TMPBUFLEN - 1)
474 		return -EINVAL;
475 
476 	if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
477 		return -EINVAL;
478 
479 	if (tr && (len < *size))
480 		*tr = *p;
481 
482 	*buf += len;
483 	*size -= len;
484 
485 	return 0;
486 }
487 
488 /**
489  * proc_put_long - converts an integer to a decimal ASCII formatted string
490  *
491  * @buf: the user buffer
492  * @size: the size of the user buffer
493  * @val: the integer to be converted
494  * @neg: sign of the number, %TRUE for negative
495  *
496  * In case of success @buf and @size are updated with the amount of bytes
497  * written.
498  */
499 static void proc_put_long(void **buf, size_t *size, unsigned long val, bool neg)
500 {
501 	int len;
502 	char tmp[TMPBUFLEN], *p = tmp;
503 
504 	sprintf(p, "%s%lu", neg ? "-" : "", val);
505 	len = strlen(tmp);
506 	if (len > *size)
507 		len = *size;
508 	memcpy(*buf, tmp, len);
509 	*size -= len;
510 	*buf += len;
511 }
512 #undef TMPBUFLEN
513 
514 static void proc_put_char(void **buf, size_t *size, char c)
515 {
516 	if (*size) {
517 		char **buffer = (char **)buf;
518 		**buffer = c;
519 
520 		(*size)--;
521 		(*buffer)++;
522 		*buf = *buffer;
523 	}
524 }
525 
526 static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
527 				 int *valp,
528 				 int write, void *data)
529 {
530 	if (write) {
531 		if (*negp) {
532 			if (*lvalp > (unsigned long) INT_MAX + 1)
533 				return -EINVAL;
534 			*valp = -*lvalp;
535 		} else {
536 			if (*lvalp > (unsigned long) INT_MAX)
537 				return -EINVAL;
538 			*valp = *lvalp;
539 		}
540 	} else {
541 		int val = *valp;
542 		if (val < 0) {
543 			*negp = true;
544 			*lvalp = -(unsigned long)val;
545 		} else {
546 			*negp = false;
547 			*lvalp = (unsigned long)val;
548 		}
549 	}
550 	return 0;
551 }
552 
553 static int do_proc_douintvec_conv(unsigned long *lvalp,
554 				  unsigned int *valp,
555 				  int write, void *data)
556 {
557 	if (write) {
558 		if (*lvalp > UINT_MAX)
559 			return -EINVAL;
560 		*valp = *lvalp;
561 	} else {
562 		unsigned int val = *valp;
563 		*lvalp = (unsigned long)val;
564 	}
565 	return 0;
566 }
567 
568 static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
569 
570 static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
571 		  int write, void *buffer,
572 		  size_t *lenp, loff_t *ppos,
573 		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
574 			      int write, void *data),
575 		  void *data)
576 {
577 	int *i, vleft, first = 1, err = 0;
578 	size_t left;
579 	char *p;
580 
581 	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
582 		*lenp = 0;
583 		return 0;
584 	}
585 
586 	i = (int *) tbl_data;
587 	vleft = table->maxlen / sizeof(*i);
588 	left = *lenp;
589 
590 	if (!conv)
591 		conv = do_proc_dointvec_conv;
592 
593 	if (write) {
594 		if (proc_first_pos_non_zero_ignore(ppos, table))
595 			goto out;
596 
597 		if (left > PAGE_SIZE - 1)
598 			left = PAGE_SIZE - 1;
599 		p = buffer;
600 	}
601 
602 	for (; left && vleft--; i++, first=0) {
603 		unsigned long lval;
604 		bool neg;
605 
606 		if (write) {
607 			left -= proc_skip_spaces(&p);
608 
609 			if (!left)
610 				break;
611 			err = proc_get_long(&p, &left, &lval, &neg,
612 					     proc_wspace_sep,
613 					     sizeof(proc_wspace_sep), NULL);
614 			if (err)
615 				break;
616 			if (conv(&neg, &lval, i, 1, data)) {
617 				err = -EINVAL;
618 				break;
619 			}
620 		} else {
621 			if (conv(&neg, &lval, i, 0, data)) {
622 				err = -EINVAL;
623 				break;
624 			}
625 			if (!first)
626 				proc_put_char(&buffer, &left, '\t');
627 			proc_put_long(&buffer, &left, lval, neg);
628 		}
629 	}
630 
631 	if (!write && !first && left && !err)
632 		proc_put_char(&buffer, &left, '\n');
633 	if (write && !err && left)
634 		left -= proc_skip_spaces(&p);
635 	if (write && first)
636 		return err ? : -EINVAL;
637 	*lenp -= left;
638 out:
639 	*ppos += *lenp;
640 	return err;
641 }
642 
643 static int do_proc_dointvec(struct ctl_table *table, int write,
644 		  void *buffer, size_t *lenp, loff_t *ppos,
645 		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
646 			      int write, void *data),
647 		  void *data)
648 {
649 	return __do_proc_dointvec(table->data, table, write,
650 			buffer, lenp, ppos, conv, data);
651 }
652 
653 static int do_proc_douintvec_w(unsigned int *tbl_data,
654 			       struct ctl_table *table,
655 			       void *buffer,
656 			       size_t *lenp, loff_t *ppos,
657 			       int (*conv)(unsigned long *lvalp,
658 					   unsigned int *valp,
659 					   int write, void *data),
660 			       void *data)
661 {
662 	unsigned long lval;
663 	int err = 0;
664 	size_t left;
665 	bool neg;
666 	char *p = buffer;
667 
668 	left = *lenp;
669 
670 	if (proc_first_pos_non_zero_ignore(ppos, table))
671 		goto bail_early;
672 
673 	if (left > PAGE_SIZE - 1)
674 		left = PAGE_SIZE - 1;
675 
676 	left -= proc_skip_spaces(&p);
677 	if (!left) {
678 		err = -EINVAL;
679 		goto out_free;
680 	}
681 
682 	err = proc_get_long(&p, &left, &lval, &neg,
683 			     proc_wspace_sep,
684 			     sizeof(proc_wspace_sep), NULL);
685 	if (err || neg) {
686 		err = -EINVAL;
687 		goto out_free;
688 	}
689 
690 	if (conv(&lval, tbl_data, 1, data)) {
691 		err = -EINVAL;
692 		goto out_free;
693 	}
694 
695 	if (!err && left)
696 		left -= proc_skip_spaces(&p);
697 
698 out_free:
699 	if (err)
700 		return -EINVAL;
701 
702 	return 0;
703 
704 	/* This is in keeping with old __do_proc_dointvec() */
705 bail_early:
706 	*ppos += *lenp;
707 	return err;
708 }
709 
710 static int do_proc_douintvec_r(unsigned int *tbl_data, void *buffer,
711 			       size_t *lenp, loff_t *ppos,
712 			       int (*conv)(unsigned long *lvalp,
713 					   unsigned int *valp,
714 					   int write, void *data),
715 			       void *data)
716 {
717 	unsigned long lval;
718 	int err = 0;
719 	size_t left;
720 
721 	left = *lenp;
722 
723 	if (conv(&lval, tbl_data, 0, data)) {
724 		err = -EINVAL;
725 		goto out;
726 	}
727 
728 	proc_put_long(&buffer, &left, lval, false);
729 	if (!left)
730 		goto out;
731 
732 	proc_put_char(&buffer, &left, '\n');
733 
734 out:
735 	*lenp -= left;
736 	*ppos += *lenp;
737 
738 	return err;
739 }
740 
741 static int __do_proc_douintvec(void *tbl_data, struct ctl_table *table,
742 			       int write, void *buffer,
743 			       size_t *lenp, loff_t *ppos,
744 			       int (*conv)(unsigned long *lvalp,
745 					   unsigned int *valp,
746 					   int write, void *data),
747 			       void *data)
748 {
749 	unsigned int *i, vleft;
750 
751 	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
752 		*lenp = 0;
753 		return 0;
754 	}
755 
756 	i = (unsigned int *) tbl_data;
757 	vleft = table->maxlen / sizeof(*i);
758 
759 	/*
760 	 * Arrays are not supported, keep this simple. *Do not* add
761 	 * support for them.
762 	 */
763 	if (vleft != 1) {
764 		*lenp = 0;
765 		return -EINVAL;
766 	}
767 
768 	if (!conv)
769 		conv = do_proc_douintvec_conv;
770 
771 	if (write)
772 		return do_proc_douintvec_w(i, table, buffer, lenp, ppos,
773 					   conv, data);
774 	return do_proc_douintvec_r(i, buffer, lenp, ppos, conv, data);
775 }
776 
777 static int do_proc_douintvec(struct ctl_table *table, int write,
778 			     void *buffer, size_t *lenp, loff_t *ppos,
779 			     int (*conv)(unsigned long *lvalp,
780 					 unsigned int *valp,
781 					 int write, void *data),
782 			     void *data)
783 {
784 	return __do_proc_douintvec(table->data, table, write,
785 				   buffer, lenp, ppos, conv, data);
786 }
787 
788 /**
789  * proc_dointvec - read a vector of integers
790  * @table: the sysctl table
791  * @write: %TRUE if this is a write to the sysctl file
792  * @buffer: the user buffer
793  * @lenp: the size of the user buffer
794  * @ppos: file position
795  *
796  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
797  * values from/to the user buffer, treated as an ASCII string.
798  *
799  * Returns 0 on success.
800  */
801 int proc_dointvec(struct ctl_table *table, int write, void *buffer,
802 		  size_t *lenp, loff_t *ppos)
803 {
804 	return do_proc_dointvec(table, write, buffer, lenp, ppos, NULL, NULL);
805 }
806 
807 #ifdef CONFIG_COMPACTION
808 static int proc_dointvec_minmax_warn_RT_change(struct ctl_table *table,
809 		int write, void *buffer, size_t *lenp, loff_t *ppos)
810 {
811 	int ret, old;
812 
813 	if (!IS_ENABLED(CONFIG_PREEMPT_RT) || !write)
814 		return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
815 
816 	old = *(int *)table->data;
817 	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
818 	if (ret)
819 		return ret;
820 	if (old != *(int *)table->data)
821 		pr_warn_once("sysctl attribute %s changed by %s[%d]\n",
822 			     table->procname, current->comm,
823 			     task_pid_nr(current));
824 	return ret;
825 }
826 #endif
827 
828 /**
829  * proc_douintvec - read a vector of unsigned integers
830  * @table: the sysctl table
831  * @write: %TRUE if this is a write to the sysctl file
832  * @buffer: the user buffer
833  * @lenp: the size of the user buffer
834  * @ppos: file position
835  *
836  * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
837  * values from/to the user buffer, treated as an ASCII string.
838  *
839  * Returns 0 on success.
840  */
841 int proc_douintvec(struct ctl_table *table, int write, void *buffer,
842 		size_t *lenp, loff_t *ppos)
843 {
844 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
845 				 do_proc_douintvec_conv, NULL);
846 }
847 
848 /*
849  * Taint values can only be increased
850  * This means we can safely use a temporary.
851  */
852 static int proc_taint(struct ctl_table *table, int write,
853 			       void *buffer, size_t *lenp, loff_t *ppos)
854 {
855 	struct ctl_table t;
856 	unsigned long tmptaint = get_taint();
857 	int err;
858 
859 	if (write && !capable(CAP_SYS_ADMIN))
860 		return -EPERM;
861 
862 	t = *table;
863 	t.data = &tmptaint;
864 	err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
865 	if (err < 0)
866 		return err;
867 
868 	if (write) {
869 		/*
870 		 * Poor man's atomic or. Not worth adding a primitive
871 		 * to everyone's atomic.h for this
872 		 */
873 		int i;
874 		for (i = 0; i < BITS_PER_LONG && tmptaint >> i; i++) {
875 			if ((tmptaint >> i) & 1)
876 				add_taint(i, LOCKDEP_STILL_OK);
877 		}
878 	}
879 
880 	return err;
881 }
882 
883 #ifdef CONFIG_PRINTK
884 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
885 				void *buffer, size_t *lenp, loff_t *ppos)
886 {
887 	if (write && !capable(CAP_SYS_ADMIN))
888 		return -EPERM;
889 
890 	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
891 }
892 #endif
893 
894 /**
895  * struct do_proc_dointvec_minmax_conv_param - proc_dointvec_minmax() range checking structure
896  * @min: pointer to minimum allowable value
897  * @max: pointer to maximum allowable value
898  *
899  * The do_proc_dointvec_minmax_conv_param structure provides the
900  * minimum and maximum values for doing range checking for those sysctl
901  * parameters that use the proc_dointvec_minmax() handler.
902  */
903 struct do_proc_dointvec_minmax_conv_param {
904 	int *min;
905 	int *max;
906 };
907 
908 static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
909 					int *valp,
910 					int write, void *data)
911 {
912 	int tmp, ret;
913 	struct do_proc_dointvec_minmax_conv_param *param = data;
914 	/*
915 	 * If writing, first do so via a temporary local int so we can
916 	 * bounds-check it before touching *valp.
917 	 */
918 	int *ip = write ? &tmp : valp;
919 
920 	ret = do_proc_dointvec_conv(negp, lvalp, ip, write, data);
921 	if (ret)
922 		return ret;
923 
924 	if (write) {
925 		if ((param->min && *param->min > tmp) ||
926 		    (param->max && *param->max < tmp))
927 			return -EINVAL;
928 		*valp = tmp;
929 	}
930 
931 	return 0;
932 }
933 
934 /**
935  * proc_dointvec_minmax - read a vector of integers with min/max values
936  * @table: the sysctl table
937  * @write: %TRUE if this is a write to the sysctl file
938  * @buffer: the user buffer
939  * @lenp: the size of the user buffer
940  * @ppos: file position
941  *
942  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
943  * values from/to the user buffer, treated as an ASCII string.
944  *
945  * This routine will ensure the values are within the range specified by
946  * table->extra1 (min) and table->extra2 (max).
947  *
948  * Returns 0 on success or -EINVAL on write when the range check fails.
949  */
950 int proc_dointvec_minmax(struct ctl_table *table, int write,
951 		  void *buffer, size_t *lenp, loff_t *ppos)
952 {
953 	struct do_proc_dointvec_minmax_conv_param param = {
954 		.min = (int *) table->extra1,
955 		.max = (int *) table->extra2,
956 	};
957 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
958 				do_proc_dointvec_minmax_conv, &param);
959 }
960 
961 /**
962  * struct do_proc_douintvec_minmax_conv_param - proc_douintvec_minmax() range checking structure
963  * @min: pointer to minimum allowable value
964  * @max: pointer to maximum allowable value
965  *
966  * The do_proc_douintvec_minmax_conv_param structure provides the
967  * minimum and maximum values for doing range checking for those sysctl
968  * parameters that use the proc_douintvec_minmax() handler.
969  */
970 struct do_proc_douintvec_minmax_conv_param {
971 	unsigned int *min;
972 	unsigned int *max;
973 };
974 
975 static int do_proc_douintvec_minmax_conv(unsigned long *lvalp,
976 					 unsigned int *valp,
977 					 int write, void *data)
978 {
979 	int ret;
980 	unsigned int tmp;
981 	struct do_proc_douintvec_minmax_conv_param *param = data;
982 	/* write via temporary local uint for bounds-checking */
983 	unsigned int *up = write ? &tmp : valp;
984 
985 	ret = do_proc_douintvec_conv(lvalp, up, write, data);
986 	if (ret)
987 		return ret;
988 
989 	if (write) {
990 		if ((param->min && *param->min > tmp) ||
991 		    (param->max && *param->max < tmp))
992 			return -ERANGE;
993 
994 		*valp = tmp;
995 	}
996 
997 	return 0;
998 }
999 
1000 /**
1001  * proc_douintvec_minmax - read a vector of unsigned ints with min/max values
1002  * @table: the sysctl table
1003  * @write: %TRUE if this is a write to the sysctl file
1004  * @buffer: the user buffer
1005  * @lenp: the size of the user buffer
1006  * @ppos: file position
1007  *
1008  * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
1009  * values from/to the user buffer, treated as an ASCII string. Negative
1010  * strings are not allowed.
1011  *
1012  * This routine will ensure the values are within the range specified by
1013  * table->extra1 (min) and table->extra2 (max). There is a final sanity
1014  * check for UINT_MAX to avoid having to support wrap around uses from
1015  * userspace.
1016  *
1017  * Returns 0 on success or -ERANGE on write when the range check fails.
1018  */
1019 int proc_douintvec_minmax(struct ctl_table *table, int write,
1020 			  void *buffer, size_t *lenp, loff_t *ppos)
1021 {
1022 	struct do_proc_douintvec_minmax_conv_param param = {
1023 		.min = (unsigned int *) table->extra1,
1024 		.max = (unsigned int *) table->extra2,
1025 	};
1026 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1027 				 do_proc_douintvec_minmax_conv, &param);
1028 }
1029 
1030 static int do_proc_dopipe_max_size_conv(unsigned long *lvalp,
1031 					unsigned int *valp,
1032 					int write, void *data)
1033 {
1034 	if (write) {
1035 		unsigned int val;
1036 
1037 		val = round_pipe_size(*lvalp);
1038 		if (val == 0)
1039 			return -EINVAL;
1040 
1041 		*valp = val;
1042 	} else {
1043 		unsigned int val = *valp;
1044 		*lvalp = (unsigned long) val;
1045 	}
1046 
1047 	return 0;
1048 }
1049 
1050 static int proc_dopipe_max_size(struct ctl_table *table, int write,
1051 				void *buffer, size_t *lenp, loff_t *ppos)
1052 {
1053 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1054 				 do_proc_dopipe_max_size_conv, NULL);
1055 }
1056 
1057 static void validate_coredump_safety(void)
1058 {
1059 #ifdef CONFIG_COREDUMP
1060 	if (suid_dumpable == SUID_DUMP_ROOT &&
1061 	    core_pattern[0] != '/' && core_pattern[0] != '|') {
1062 		printk(KERN_WARNING
1063 "Unsafe core_pattern used with fs.suid_dumpable=2.\n"
1064 "Pipe handler or fully qualified core dump path required.\n"
1065 "Set kernel.core_pattern before fs.suid_dumpable.\n"
1066 		);
1067 	}
1068 #endif
1069 }
1070 
1071 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
1072 		void *buffer, size_t *lenp, loff_t *ppos)
1073 {
1074 	int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
1075 	if (!error)
1076 		validate_coredump_safety();
1077 	return error;
1078 }
1079 
1080 #ifdef CONFIG_COREDUMP
1081 static int proc_dostring_coredump(struct ctl_table *table, int write,
1082 		  void *buffer, size_t *lenp, loff_t *ppos)
1083 {
1084 	int error = proc_dostring(table, write, buffer, lenp, ppos);
1085 	if (!error)
1086 		validate_coredump_safety();
1087 	return error;
1088 }
1089 #endif
1090 
1091 #ifdef CONFIG_MAGIC_SYSRQ
1092 static int sysrq_sysctl_handler(struct ctl_table *table, int write,
1093 				void *buffer, size_t *lenp, loff_t *ppos)
1094 {
1095 	int tmp, ret;
1096 
1097 	tmp = sysrq_mask();
1098 
1099 	ret = __do_proc_dointvec(&tmp, table, write, buffer,
1100 			       lenp, ppos, NULL, NULL);
1101 	if (ret || !write)
1102 		return ret;
1103 
1104 	if (write)
1105 		sysrq_toggle_support(tmp);
1106 
1107 	return 0;
1108 }
1109 #endif
1110 
1111 static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table,
1112 		int write, void *buffer, size_t *lenp, loff_t *ppos,
1113 		unsigned long convmul, unsigned long convdiv)
1114 {
1115 	unsigned long *i, *min, *max;
1116 	int vleft, first = 1, err = 0;
1117 	size_t left;
1118 	char *p;
1119 
1120 	if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
1121 		*lenp = 0;
1122 		return 0;
1123 	}
1124 
1125 	i = (unsigned long *) data;
1126 	min = (unsigned long *) table->extra1;
1127 	max = (unsigned long *) table->extra2;
1128 	vleft = table->maxlen / sizeof(unsigned long);
1129 	left = *lenp;
1130 
1131 	if (write) {
1132 		if (proc_first_pos_non_zero_ignore(ppos, table))
1133 			goto out;
1134 
1135 		if (left > PAGE_SIZE - 1)
1136 			left = PAGE_SIZE - 1;
1137 		p = buffer;
1138 	}
1139 
1140 	for (; left && vleft--; i++, first = 0) {
1141 		unsigned long val;
1142 
1143 		if (write) {
1144 			bool neg;
1145 
1146 			left -= proc_skip_spaces(&p);
1147 			if (!left)
1148 				break;
1149 
1150 			err = proc_get_long(&p, &left, &val, &neg,
1151 					     proc_wspace_sep,
1152 					     sizeof(proc_wspace_sep), NULL);
1153 			if (err)
1154 				break;
1155 			if (neg)
1156 				continue;
1157 			val = convmul * val / convdiv;
1158 			if ((min && val < *min) || (max && val > *max)) {
1159 				err = -EINVAL;
1160 				break;
1161 			}
1162 			*i = val;
1163 		} else {
1164 			val = convdiv * (*i) / convmul;
1165 			if (!first)
1166 				proc_put_char(&buffer, &left, '\t');
1167 			proc_put_long(&buffer, &left, val, false);
1168 		}
1169 	}
1170 
1171 	if (!write && !first && left && !err)
1172 		proc_put_char(&buffer, &left, '\n');
1173 	if (write && !err)
1174 		left -= proc_skip_spaces(&p);
1175 	if (write && first)
1176 		return err ? : -EINVAL;
1177 	*lenp -= left;
1178 out:
1179 	*ppos += *lenp;
1180 	return err;
1181 }
1182 
1183 static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
1184 		void *buffer, size_t *lenp, loff_t *ppos, unsigned long convmul,
1185 		unsigned long convdiv)
1186 {
1187 	return __do_proc_doulongvec_minmax(table->data, table, write,
1188 			buffer, lenp, ppos, convmul, convdiv);
1189 }
1190 
1191 /**
1192  * proc_doulongvec_minmax - read a vector of long integers with min/max values
1193  * @table: the sysctl table
1194  * @write: %TRUE if this is a write to the sysctl file
1195  * @buffer: the user buffer
1196  * @lenp: the size of the user buffer
1197  * @ppos: file position
1198  *
1199  * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1200  * values from/to the user buffer, treated as an ASCII string.
1201  *
1202  * This routine will ensure the values are within the range specified by
1203  * table->extra1 (min) and table->extra2 (max).
1204  *
1205  * Returns 0 on success.
1206  */
1207 int proc_doulongvec_minmax(struct ctl_table *table, int write,
1208 			   void *buffer, size_t *lenp, loff_t *ppos)
1209 {
1210     return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
1211 }
1212 
1213 /**
1214  * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
1215  * @table: the sysctl table
1216  * @write: %TRUE if this is a write to the sysctl file
1217  * @buffer: the user buffer
1218  * @lenp: the size of the user buffer
1219  * @ppos: file position
1220  *
1221  * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1222  * values from/to the user buffer, treated as an ASCII string. The values
1223  * are treated as milliseconds, and converted to jiffies when they are stored.
1224  *
1225  * This routine will ensure the values are within the range specified by
1226  * table->extra1 (min) and table->extra2 (max).
1227  *
1228  * Returns 0 on success.
1229  */
1230 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1231 				      void *buffer, size_t *lenp, loff_t *ppos)
1232 {
1233     return do_proc_doulongvec_minmax(table, write, buffer,
1234 				     lenp, ppos, HZ, 1000l);
1235 }
1236 
1237 
1238 static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
1239 					 int *valp,
1240 					 int write, void *data)
1241 {
1242 	if (write) {
1243 		if (*lvalp > INT_MAX / HZ)
1244 			return 1;
1245 		*valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
1246 	} else {
1247 		int val = *valp;
1248 		unsigned long lval;
1249 		if (val < 0) {
1250 			*negp = true;
1251 			lval = -(unsigned long)val;
1252 		} else {
1253 			*negp = false;
1254 			lval = (unsigned long)val;
1255 		}
1256 		*lvalp = lval / HZ;
1257 	}
1258 	return 0;
1259 }
1260 
1261 static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
1262 						int *valp,
1263 						int write, void *data)
1264 {
1265 	if (write) {
1266 		if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
1267 			return 1;
1268 		*valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
1269 	} else {
1270 		int val = *valp;
1271 		unsigned long lval;
1272 		if (val < 0) {
1273 			*negp = true;
1274 			lval = -(unsigned long)val;
1275 		} else {
1276 			*negp = false;
1277 			lval = (unsigned long)val;
1278 		}
1279 		*lvalp = jiffies_to_clock_t(lval);
1280 	}
1281 	return 0;
1282 }
1283 
1284 static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
1285 					    int *valp,
1286 					    int write, void *data)
1287 {
1288 	if (write) {
1289 		unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
1290 
1291 		if (jif > INT_MAX)
1292 			return 1;
1293 		*valp = (int)jif;
1294 	} else {
1295 		int val = *valp;
1296 		unsigned long lval;
1297 		if (val < 0) {
1298 			*negp = true;
1299 			lval = -(unsigned long)val;
1300 		} else {
1301 			*negp = false;
1302 			lval = (unsigned long)val;
1303 		}
1304 		*lvalp = jiffies_to_msecs(lval);
1305 	}
1306 	return 0;
1307 }
1308 
1309 /**
1310  * proc_dointvec_jiffies - read a vector of integers as seconds
1311  * @table: the sysctl table
1312  * @write: %TRUE if this is a write to the sysctl file
1313  * @buffer: the user buffer
1314  * @lenp: the size of the user buffer
1315  * @ppos: file position
1316  *
1317  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1318  * values from/to the user buffer, treated as an ASCII string.
1319  * The values read are assumed to be in seconds, and are converted into
1320  * jiffies.
1321  *
1322  * Returns 0 on success.
1323  */
1324 int proc_dointvec_jiffies(struct ctl_table *table, int write,
1325 			  void *buffer, size_t *lenp, loff_t *ppos)
1326 {
1327     return do_proc_dointvec(table,write,buffer,lenp,ppos,
1328 		    	    do_proc_dointvec_jiffies_conv,NULL);
1329 }
1330 
1331 /**
1332  * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
1333  * @table: the sysctl table
1334  * @write: %TRUE if this is a write to the sysctl file
1335  * @buffer: the user buffer
1336  * @lenp: the size of the user buffer
1337  * @ppos: pointer to the file position
1338  *
1339  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1340  * values from/to the user buffer, treated as an ASCII string.
1341  * The values read are assumed to be in 1/USER_HZ seconds, and
1342  * are converted into jiffies.
1343  *
1344  * Returns 0 on success.
1345  */
1346 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1347 				 void *buffer, size_t *lenp, loff_t *ppos)
1348 {
1349     return do_proc_dointvec(table,write,buffer,lenp,ppos,
1350 		    	    do_proc_dointvec_userhz_jiffies_conv,NULL);
1351 }
1352 
1353 /**
1354  * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
1355  * @table: the sysctl table
1356  * @write: %TRUE if this is a write to the sysctl file
1357  * @buffer: the user buffer
1358  * @lenp: the size of the user buffer
1359  * @ppos: file position
1360  * @ppos: the current position in the file
1361  *
1362  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1363  * values from/to the user buffer, treated as an ASCII string.
1364  * The values read are assumed to be in 1/1000 seconds, and
1365  * are converted into jiffies.
1366  *
1367  * Returns 0 on success.
1368  */
1369 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write, void *buffer,
1370 		size_t *lenp, loff_t *ppos)
1371 {
1372 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
1373 				do_proc_dointvec_ms_jiffies_conv, NULL);
1374 }
1375 
1376 static int proc_do_cad_pid(struct ctl_table *table, int write, void *buffer,
1377 		size_t *lenp, loff_t *ppos)
1378 {
1379 	struct pid *new_pid;
1380 	pid_t tmp;
1381 	int r;
1382 
1383 	tmp = pid_vnr(cad_pid);
1384 
1385 	r = __do_proc_dointvec(&tmp, table, write, buffer,
1386 			       lenp, ppos, NULL, NULL);
1387 	if (r || !write)
1388 		return r;
1389 
1390 	new_pid = find_get_pid(tmp);
1391 	if (!new_pid)
1392 		return -ESRCH;
1393 
1394 	put_pid(xchg(&cad_pid, new_pid));
1395 	return 0;
1396 }
1397 
1398 /**
1399  * proc_do_large_bitmap - read/write from/to a large bitmap
1400  * @table: the sysctl table
1401  * @write: %TRUE if this is a write to the sysctl file
1402  * @buffer: the user buffer
1403  * @lenp: the size of the user buffer
1404  * @ppos: file position
1405  *
1406  * The bitmap is stored at table->data and the bitmap length (in bits)
1407  * in table->maxlen.
1408  *
1409  * We use a range comma separated format (e.g. 1,3-4,10-10) so that
1410  * large bitmaps may be represented in a compact manner. Writing into
1411  * the file will clear the bitmap then update it with the given input.
1412  *
1413  * Returns 0 on success.
1414  */
1415 int proc_do_large_bitmap(struct ctl_table *table, int write,
1416 			 void *buffer, size_t *lenp, loff_t *ppos)
1417 {
1418 	int err = 0;
1419 	bool first = 1;
1420 	size_t left = *lenp;
1421 	unsigned long bitmap_len = table->maxlen;
1422 	unsigned long *bitmap = *(unsigned long **) table->data;
1423 	unsigned long *tmp_bitmap = NULL;
1424 	char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
1425 
1426 	if (!bitmap || !bitmap_len || !left || (*ppos && !write)) {
1427 		*lenp = 0;
1428 		return 0;
1429 	}
1430 
1431 	if (write) {
1432 		char *p = buffer;
1433 		size_t skipped = 0;
1434 
1435 		if (left > PAGE_SIZE - 1) {
1436 			left = PAGE_SIZE - 1;
1437 			/* How much of the buffer we'll skip this pass */
1438 			skipped = *lenp - left;
1439 		}
1440 
1441 		tmp_bitmap = bitmap_zalloc(bitmap_len, GFP_KERNEL);
1442 		if (!tmp_bitmap)
1443 			return -ENOMEM;
1444 		proc_skip_char(&p, &left, '\n');
1445 		while (!err && left) {
1446 			unsigned long val_a, val_b;
1447 			bool neg;
1448 			size_t saved_left;
1449 
1450 			/* In case we stop parsing mid-number, we can reset */
1451 			saved_left = left;
1452 			err = proc_get_long(&p, &left, &val_a, &neg, tr_a,
1453 					     sizeof(tr_a), &c);
1454 			/*
1455 			 * If we consumed the entirety of a truncated buffer or
1456 			 * only one char is left (may be a "-"), then stop here,
1457 			 * reset, & come back for more.
1458 			 */
1459 			if ((left <= 1) && skipped) {
1460 				left = saved_left;
1461 				break;
1462 			}
1463 
1464 			if (err)
1465 				break;
1466 			if (val_a >= bitmap_len || neg) {
1467 				err = -EINVAL;
1468 				break;
1469 			}
1470 
1471 			val_b = val_a;
1472 			if (left) {
1473 				p++;
1474 				left--;
1475 			}
1476 
1477 			if (c == '-') {
1478 				err = proc_get_long(&p, &left, &val_b,
1479 						     &neg, tr_b, sizeof(tr_b),
1480 						     &c);
1481 				/*
1482 				 * If we consumed all of a truncated buffer or
1483 				 * then stop here, reset, & come back for more.
1484 				 */
1485 				if (!left && skipped) {
1486 					left = saved_left;
1487 					break;
1488 				}
1489 
1490 				if (err)
1491 					break;
1492 				if (val_b >= bitmap_len || neg ||
1493 				    val_a > val_b) {
1494 					err = -EINVAL;
1495 					break;
1496 				}
1497 				if (left) {
1498 					p++;
1499 					left--;
1500 				}
1501 			}
1502 
1503 			bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
1504 			first = 0;
1505 			proc_skip_char(&p, &left, '\n');
1506 		}
1507 		left += skipped;
1508 	} else {
1509 		unsigned long bit_a, bit_b = 0;
1510 
1511 		while (left) {
1512 			bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
1513 			if (bit_a >= bitmap_len)
1514 				break;
1515 			bit_b = find_next_zero_bit(bitmap, bitmap_len,
1516 						   bit_a + 1) - 1;
1517 
1518 			if (!first)
1519 				proc_put_char(&buffer, &left, ',');
1520 			proc_put_long(&buffer, &left, bit_a, false);
1521 			if (bit_a != bit_b) {
1522 				proc_put_char(&buffer, &left, '-');
1523 				proc_put_long(&buffer, &left, bit_b, false);
1524 			}
1525 
1526 			first = 0; bit_b++;
1527 		}
1528 		proc_put_char(&buffer, &left, '\n');
1529 	}
1530 
1531 	if (!err) {
1532 		if (write) {
1533 			if (*ppos)
1534 				bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
1535 			else
1536 				bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
1537 		}
1538 		*lenp -= left;
1539 		*ppos += *lenp;
1540 	}
1541 
1542 	bitmap_free(tmp_bitmap);
1543 	return err;
1544 }
1545 
1546 #else /* CONFIG_PROC_SYSCTL */
1547 
1548 int proc_dostring(struct ctl_table *table, int write,
1549 		  void *buffer, size_t *lenp, loff_t *ppos)
1550 {
1551 	return -ENOSYS;
1552 }
1553 
1554 int proc_dointvec(struct ctl_table *table, int write,
1555 		  void *buffer, size_t *lenp, loff_t *ppos)
1556 {
1557 	return -ENOSYS;
1558 }
1559 
1560 int proc_douintvec(struct ctl_table *table, int write,
1561 		  void *buffer, size_t *lenp, loff_t *ppos)
1562 {
1563 	return -ENOSYS;
1564 }
1565 
1566 int proc_dointvec_minmax(struct ctl_table *table, int write,
1567 		    void *buffer, size_t *lenp, loff_t *ppos)
1568 {
1569 	return -ENOSYS;
1570 }
1571 
1572 int proc_douintvec_minmax(struct ctl_table *table, int write,
1573 			  void *buffer, size_t *lenp, loff_t *ppos)
1574 {
1575 	return -ENOSYS;
1576 }
1577 
1578 int proc_dointvec_jiffies(struct ctl_table *table, int write,
1579 		    void *buffer, size_t *lenp, loff_t *ppos)
1580 {
1581 	return -ENOSYS;
1582 }
1583 
1584 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1585 		    void *buffer, size_t *lenp, loff_t *ppos)
1586 {
1587 	return -ENOSYS;
1588 }
1589 
1590 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
1591 			     void *buffer, size_t *lenp, loff_t *ppos)
1592 {
1593 	return -ENOSYS;
1594 }
1595 
1596 int proc_doulongvec_minmax(struct ctl_table *table, int write,
1597 		    void *buffer, size_t *lenp, loff_t *ppos)
1598 {
1599 	return -ENOSYS;
1600 }
1601 
1602 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1603 				      void *buffer, size_t *lenp, loff_t *ppos)
1604 {
1605 	return -ENOSYS;
1606 }
1607 
1608 int proc_do_large_bitmap(struct ctl_table *table, int write,
1609 			 void *buffer, size_t *lenp, loff_t *ppos)
1610 {
1611 	return -ENOSYS;
1612 }
1613 
1614 #endif /* CONFIG_PROC_SYSCTL */
1615 
1616 #if defined(CONFIG_SYSCTL)
1617 int proc_do_static_key(struct ctl_table *table, int write,
1618 		       void *buffer, size_t *lenp, loff_t *ppos)
1619 {
1620 	struct static_key *key = (struct static_key *)table->data;
1621 	static DEFINE_MUTEX(static_key_mutex);
1622 	int val, ret;
1623 	struct ctl_table tmp = {
1624 		.data   = &val,
1625 		.maxlen = sizeof(val),
1626 		.mode   = table->mode,
1627 		.extra1 = SYSCTL_ZERO,
1628 		.extra2 = SYSCTL_ONE,
1629 	};
1630 
1631 	if (write && !capable(CAP_SYS_ADMIN))
1632 		return -EPERM;
1633 
1634 	mutex_lock(&static_key_mutex);
1635 	val = static_key_enabled(key);
1636 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
1637 	if (write && !ret) {
1638 		if (val)
1639 			static_key_enable(key);
1640 		else
1641 			static_key_disable(key);
1642 	}
1643 	mutex_unlock(&static_key_mutex);
1644 	return ret;
1645 }
1646 
1647 static struct ctl_table kern_table[] = {
1648 	{
1649 		.procname	= "sched_child_runs_first",
1650 		.data		= &sysctl_sched_child_runs_first,
1651 		.maxlen		= sizeof(unsigned int),
1652 		.mode		= 0644,
1653 		.proc_handler	= proc_dointvec,
1654 	},
1655 #ifdef CONFIG_SCHED_DEBUG
1656 	{
1657 		.procname	= "sched_min_granularity_ns",
1658 		.data		= &sysctl_sched_min_granularity,
1659 		.maxlen		= sizeof(unsigned int),
1660 		.mode		= 0644,
1661 		.proc_handler	= sched_proc_update_handler,
1662 		.extra1		= &min_sched_granularity_ns,
1663 		.extra2		= &max_sched_granularity_ns,
1664 	},
1665 	{
1666 		.procname	= "sched_latency_ns",
1667 		.data		= &sysctl_sched_latency,
1668 		.maxlen		= sizeof(unsigned int),
1669 		.mode		= 0644,
1670 		.proc_handler	= sched_proc_update_handler,
1671 		.extra1		= &min_sched_granularity_ns,
1672 		.extra2		= &max_sched_granularity_ns,
1673 	},
1674 	{
1675 		.procname	= "sched_wakeup_granularity_ns",
1676 		.data		= &sysctl_sched_wakeup_granularity,
1677 		.maxlen		= sizeof(unsigned int),
1678 		.mode		= 0644,
1679 		.proc_handler	= sched_proc_update_handler,
1680 		.extra1		= &min_wakeup_granularity_ns,
1681 		.extra2		= &max_wakeup_granularity_ns,
1682 	},
1683 #ifdef CONFIG_SMP
1684 	{
1685 		.procname	= "sched_tunable_scaling",
1686 		.data		= &sysctl_sched_tunable_scaling,
1687 		.maxlen		= sizeof(enum sched_tunable_scaling),
1688 		.mode		= 0644,
1689 		.proc_handler	= sched_proc_update_handler,
1690 		.extra1		= &min_sched_tunable_scaling,
1691 		.extra2		= &max_sched_tunable_scaling,
1692 	},
1693 	{
1694 		.procname	= "sched_migration_cost_ns",
1695 		.data		= &sysctl_sched_migration_cost,
1696 		.maxlen		= sizeof(unsigned int),
1697 		.mode		= 0644,
1698 		.proc_handler	= proc_dointvec,
1699 	},
1700 	{
1701 		.procname	= "sched_nr_migrate",
1702 		.data		= &sysctl_sched_nr_migrate,
1703 		.maxlen		= sizeof(unsigned int),
1704 		.mode		= 0644,
1705 		.proc_handler	= proc_dointvec,
1706 	},
1707 #ifdef CONFIG_SCHEDSTATS
1708 	{
1709 		.procname	= "sched_schedstats",
1710 		.data		= NULL,
1711 		.maxlen		= sizeof(unsigned int),
1712 		.mode		= 0644,
1713 		.proc_handler	= sysctl_schedstats,
1714 		.extra1		= SYSCTL_ZERO,
1715 		.extra2		= SYSCTL_ONE,
1716 	},
1717 #endif /* CONFIG_SCHEDSTATS */
1718 #endif /* CONFIG_SMP */
1719 #ifdef CONFIG_NUMA_BALANCING
1720 	{
1721 		.procname	= "numa_balancing_scan_delay_ms",
1722 		.data		= &sysctl_numa_balancing_scan_delay,
1723 		.maxlen		= sizeof(unsigned int),
1724 		.mode		= 0644,
1725 		.proc_handler	= proc_dointvec,
1726 	},
1727 	{
1728 		.procname	= "numa_balancing_scan_period_min_ms",
1729 		.data		= &sysctl_numa_balancing_scan_period_min,
1730 		.maxlen		= sizeof(unsigned int),
1731 		.mode		= 0644,
1732 		.proc_handler	= proc_dointvec,
1733 	},
1734 	{
1735 		.procname	= "numa_balancing_scan_period_max_ms",
1736 		.data		= &sysctl_numa_balancing_scan_period_max,
1737 		.maxlen		= sizeof(unsigned int),
1738 		.mode		= 0644,
1739 		.proc_handler	= proc_dointvec,
1740 	},
1741 	{
1742 		.procname	= "numa_balancing_scan_size_mb",
1743 		.data		= &sysctl_numa_balancing_scan_size,
1744 		.maxlen		= sizeof(unsigned int),
1745 		.mode		= 0644,
1746 		.proc_handler	= proc_dointvec_minmax,
1747 		.extra1		= SYSCTL_ONE,
1748 	},
1749 	{
1750 		.procname	= "numa_balancing",
1751 		.data		= NULL, /* filled in by handler */
1752 		.maxlen		= sizeof(unsigned int),
1753 		.mode		= 0644,
1754 		.proc_handler	= sysctl_numa_balancing,
1755 		.extra1		= SYSCTL_ZERO,
1756 		.extra2		= SYSCTL_ONE,
1757 	},
1758 #endif /* CONFIG_NUMA_BALANCING */
1759 #endif /* CONFIG_SCHED_DEBUG */
1760 	{
1761 		.procname	= "sched_rt_period_us",
1762 		.data		= &sysctl_sched_rt_period,
1763 		.maxlen		= sizeof(unsigned int),
1764 		.mode		= 0644,
1765 		.proc_handler	= sched_rt_handler,
1766 	},
1767 	{
1768 		.procname	= "sched_rt_runtime_us",
1769 		.data		= &sysctl_sched_rt_runtime,
1770 		.maxlen		= sizeof(int),
1771 		.mode		= 0644,
1772 		.proc_handler	= sched_rt_handler,
1773 	},
1774 	{
1775 		.procname	= "sched_rr_timeslice_ms",
1776 		.data		= &sysctl_sched_rr_timeslice,
1777 		.maxlen		= sizeof(int),
1778 		.mode		= 0644,
1779 		.proc_handler	= sched_rr_handler,
1780 	},
1781 #ifdef CONFIG_UCLAMP_TASK
1782 	{
1783 		.procname	= "sched_util_clamp_min",
1784 		.data		= &sysctl_sched_uclamp_util_min,
1785 		.maxlen		= sizeof(unsigned int),
1786 		.mode		= 0644,
1787 		.proc_handler	= sysctl_sched_uclamp_handler,
1788 	},
1789 	{
1790 		.procname	= "sched_util_clamp_max",
1791 		.data		= &sysctl_sched_uclamp_util_max,
1792 		.maxlen		= sizeof(unsigned int),
1793 		.mode		= 0644,
1794 		.proc_handler	= sysctl_sched_uclamp_handler,
1795 	},
1796 #endif
1797 #ifdef CONFIG_SCHED_AUTOGROUP
1798 	{
1799 		.procname	= "sched_autogroup_enabled",
1800 		.data		= &sysctl_sched_autogroup_enabled,
1801 		.maxlen		= sizeof(unsigned int),
1802 		.mode		= 0644,
1803 		.proc_handler	= proc_dointvec_minmax,
1804 		.extra1		= SYSCTL_ZERO,
1805 		.extra2		= SYSCTL_ONE,
1806 	},
1807 #endif
1808 #ifdef CONFIG_CFS_BANDWIDTH
1809 	{
1810 		.procname	= "sched_cfs_bandwidth_slice_us",
1811 		.data		= &sysctl_sched_cfs_bandwidth_slice,
1812 		.maxlen		= sizeof(unsigned int),
1813 		.mode		= 0644,
1814 		.proc_handler	= proc_dointvec_minmax,
1815 		.extra1		= SYSCTL_ONE,
1816 	},
1817 #endif
1818 #if defined(CONFIG_ENERGY_MODEL) && defined(CONFIG_CPU_FREQ_GOV_SCHEDUTIL)
1819 	{
1820 		.procname	= "sched_energy_aware",
1821 		.data		= &sysctl_sched_energy_aware,
1822 		.maxlen		= sizeof(unsigned int),
1823 		.mode		= 0644,
1824 		.proc_handler	= sched_energy_aware_handler,
1825 		.extra1		= SYSCTL_ZERO,
1826 		.extra2		= SYSCTL_ONE,
1827 	},
1828 #endif
1829 #ifdef CONFIG_PROVE_LOCKING
1830 	{
1831 		.procname	= "prove_locking",
1832 		.data		= &prove_locking,
1833 		.maxlen		= sizeof(int),
1834 		.mode		= 0644,
1835 		.proc_handler	= proc_dointvec,
1836 	},
1837 #endif
1838 #ifdef CONFIG_LOCK_STAT
1839 	{
1840 		.procname	= "lock_stat",
1841 		.data		= &lock_stat,
1842 		.maxlen		= sizeof(int),
1843 		.mode		= 0644,
1844 		.proc_handler	= proc_dointvec,
1845 	},
1846 #endif
1847 	{
1848 		.procname	= "panic",
1849 		.data		= &panic_timeout,
1850 		.maxlen		= sizeof(int),
1851 		.mode		= 0644,
1852 		.proc_handler	= proc_dointvec,
1853 	},
1854 #ifdef CONFIG_COREDUMP
1855 	{
1856 		.procname	= "core_uses_pid",
1857 		.data		= &core_uses_pid,
1858 		.maxlen		= sizeof(int),
1859 		.mode		= 0644,
1860 		.proc_handler	= proc_dointvec,
1861 	},
1862 	{
1863 		.procname	= "core_pattern",
1864 		.data		= core_pattern,
1865 		.maxlen		= CORENAME_MAX_SIZE,
1866 		.mode		= 0644,
1867 		.proc_handler	= proc_dostring_coredump,
1868 	},
1869 	{
1870 		.procname	= "core_pipe_limit",
1871 		.data		= &core_pipe_limit,
1872 		.maxlen		= sizeof(unsigned int),
1873 		.mode		= 0644,
1874 		.proc_handler	= proc_dointvec,
1875 	},
1876 #endif
1877 #ifdef CONFIG_PROC_SYSCTL
1878 	{
1879 		.procname	= "tainted",
1880 		.maxlen 	= sizeof(long),
1881 		.mode		= 0644,
1882 		.proc_handler	= proc_taint,
1883 	},
1884 	{
1885 		.procname	= "sysctl_writes_strict",
1886 		.data		= &sysctl_writes_strict,
1887 		.maxlen		= sizeof(int),
1888 		.mode		= 0644,
1889 		.proc_handler	= proc_dointvec_minmax,
1890 		.extra1		= &neg_one,
1891 		.extra2		= SYSCTL_ONE,
1892 	},
1893 #endif
1894 #ifdef CONFIG_LATENCYTOP
1895 	{
1896 		.procname	= "latencytop",
1897 		.data		= &latencytop_enabled,
1898 		.maxlen		= sizeof(int),
1899 		.mode		= 0644,
1900 		.proc_handler	= sysctl_latencytop,
1901 	},
1902 #endif
1903 #ifdef CONFIG_BLK_DEV_INITRD
1904 	{
1905 		.procname	= "real-root-dev",
1906 		.data		= &real_root_dev,
1907 		.maxlen		= sizeof(int),
1908 		.mode		= 0644,
1909 		.proc_handler	= proc_dointvec,
1910 	},
1911 #endif
1912 	{
1913 		.procname	= "print-fatal-signals",
1914 		.data		= &print_fatal_signals,
1915 		.maxlen		= sizeof(int),
1916 		.mode		= 0644,
1917 		.proc_handler	= proc_dointvec,
1918 	},
1919 #ifdef CONFIG_SPARC
1920 	{
1921 		.procname	= "reboot-cmd",
1922 		.data		= reboot_command,
1923 		.maxlen		= 256,
1924 		.mode		= 0644,
1925 		.proc_handler	= proc_dostring,
1926 	},
1927 	{
1928 		.procname	= "stop-a",
1929 		.data		= &stop_a_enabled,
1930 		.maxlen		= sizeof (int),
1931 		.mode		= 0644,
1932 		.proc_handler	= proc_dointvec,
1933 	},
1934 	{
1935 		.procname	= "scons-poweroff",
1936 		.data		= &scons_pwroff,
1937 		.maxlen		= sizeof (int),
1938 		.mode		= 0644,
1939 		.proc_handler	= proc_dointvec,
1940 	},
1941 #endif
1942 #ifdef CONFIG_SPARC64
1943 	{
1944 		.procname	= "tsb-ratio",
1945 		.data		= &sysctl_tsb_ratio,
1946 		.maxlen		= sizeof (int),
1947 		.mode		= 0644,
1948 		.proc_handler	= proc_dointvec,
1949 	},
1950 #endif
1951 #ifdef CONFIG_PARISC
1952 	{
1953 		.procname	= "soft-power",
1954 		.data		= &pwrsw_enabled,
1955 		.maxlen		= sizeof (int),
1956 		.mode		= 0644,
1957 		.proc_handler	= proc_dointvec,
1958 	},
1959 #endif
1960 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
1961 	{
1962 		.procname	= "unaligned-trap",
1963 		.data		= &unaligned_enabled,
1964 		.maxlen		= sizeof (int),
1965 		.mode		= 0644,
1966 		.proc_handler	= proc_dointvec,
1967 	},
1968 #endif
1969 	{
1970 		.procname	= "ctrl-alt-del",
1971 		.data		= &C_A_D,
1972 		.maxlen		= sizeof(int),
1973 		.mode		= 0644,
1974 		.proc_handler	= proc_dointvec,
1975 	},
1976 #ifdef CONFIG_FUNCTION_TRACER
1977 	{
1978 		.procname	= "ftrace_enabled",
1979 		.data		= &ftrace_enabled,
1980 		.maxlen		= sizeof(int),
1981 		.mode		= 0644,
1982 		.proc_handler	= ftrace_enable_sysctl,
1983 	},
1984 #endif
1985 #ifdef CONFIG_STACK_TRACER
1986 	{
1987 		.procname	= "stack_tracer_enabled",
1988 		.data		= &stack_tracer_enabled,
1989 		.maxlen		= sizeof(int),
1990 		.mode		= 0644,
1991 		.proc_handler	= stack_trace_sysctl,
1992 	},
1993 #endif
1994 #ifdef CONFIG_TRACING
1995 	{
1996 		.procname	= "ftrace_dump_on_oops",
1997 		.data		= &ftrace_dump_on_oops,
1998 		.maxlen		= sizeof(int),
1999 		.mode		= 0644,
2000 		.proc_handler	= proc_dointvec,
2001 	},
2002 	{
2003 		.procname	= "traceoff_on_warning",
2004 		.data		= &__disable_trace_on_warning,
2005 		.maxlen		= sizeof(__disable_trace_on_warning),
2006 		.mode		= 0644,
2007 		.proc_handler	= proc_dointvec,
2008 	},
2009 	{
2010 		.procname	= "tracepoint_printk",
2011 		.data		= &tracepoint_printk,
2012 		.maxlen		= sizeof(tracepoint_printk),
2013 		.mode		= 0644,
2014 		.proc_handler	= tracepoint_printk_sysctl,
2015 	},
2016 #endif
2017 #ifdef CONFIG_KEXEC_CORE
2018 	{
2019 		.procname	= "kexec_load_disabled",
2020 		.data		= &kexec_load_disabled,
2021 		.maxlen		= sizeof(int),
2022 		.mode		= 0644,
2023 		/* only handle a transition from default "0" to "1" */
2024 		.proc_handler	= proc_dointvec_minmax,
2025 		.extra1		= SYSCTL_ONE,
2026 		.extra2		= SYSCTL_ONE,
2027 	},
2028 #endif
2029 #ifdef CONFIG_MODULES
2030 	{
2031 		.procname	= "modprobe",
2032 		.data		= &modprobe_path,
2033 		.maxlen		= KMOD_PATH_LEN,
2034 		.mode		= 0644,
2035 		.proc_handler	= proc_dostring,
2036 	},
2037 	{
2038 		.procname	= "modules_disabled",
2039 		.data		= &modules_disabled,
2040 		.maxlen		= sizeof(int),
2041 		.mode		= 0644,
2042 		/* only handle a transition from default "0" to "1" */
2043 		.proc_handler	= proc_dointvec_minmax,
2044 		.extra1		= SYSCTL_ONE,
2045 		.extra2		= SYSCTL_ONE,
2046 	},
2047 #endif
2048 #ifdef CONFIG_UEVENT_HELPER
2049 	{
2050 		.procname	= "hotplug",
2051 		.data		= &uevent_helper,
2052 		.maxlen		= UEVENT_HELPER_PATH_LEN,
2053 		.mode		= 0644,
2054 		.proc_handler	= proc_dostring,
2055 	},
2056 #endif
2057 #ifdef CONFIG_CHR_DEV_SG
2058 	{
2059 		.procname	= "sg-big-buff",
2060 		.data		= &sg_big_buff,
2061 		.maxlen		= sizeof (int),
2062 		.mode		= 0444,
2063 		.proc_handler	= proc_dointvec,
2064 	},
2065 #endif
2066 #ifdef CONFIG_BSD_PROCESS_ACCT
2067 	{
2068 		.procname	= "acct",
2069 		.data		= &acct_parm,
2070 		.maxlen		= 3*sizeof(int),
2071 		.mode		= 0644,
2072 		.proc_handler	= proc_dointvec,
2073 	},
2074 #endif
2075 #ifdef CONFIG_MAGIC_SYSRQ
2076 	{
2077 		.procname	= "sysrq",
2078 		.data		= NULL,
2079 		.maxlen		= sizeof (int),
2080 		.mode		= 0644,
2081 		.proc_handler	= sysrq_sysctl_handler,
2082 	},
2083 #endif
2084 #ifdef CONFIG_PROC_SYSCTL
2085 	{
2086 		.procname	= "cad_pid",
2087 		.data		= NULL,
2088 		.maxlen		= sizeof (int),
2089 		.mode		= 0600,
2090 		.proc_handler	= proc_do_cad_pid,
2091 	},
2092 #endif
2093 	{
2094 		.procname	= "threads-max",
2095 		.data		= NULL,
2096 		.maxlen		= sizeof(int),
2097 		.mode		= 0644,
2098 		.proc_handler	= sysctl_max_threads,
2099 	},
2100 	{
2101 		.procname	= "random",
2102 		.mode		= 0555,
2103 		.child		= random_table,
2104 	},
2105 	{
2106 		.procname	= "usermodehelper",
2107 		.mode		= 0555,
2108 		.child		= usermodehelper_table,
2109 	},
2110 #ifdef CONFIG_FW_LOADER_USER_HELPER
2111 	{
2112 		.procname	= "firmware_config",
2113 		.mode		= 0555,
2114 		.child		= firmware_config_table,
2115 	},
2116 #endif
2117 	{
2118 		.procname	= "overflowuid",
2119 		.data		= &overflowuid,
2120 		.maxlen		= sizeof(int),
2121 		.mode		= 0644,
2122 		.proc_handler	= proc_dointvec_minmax,
2123 		.extra1		= &minolduid,
2124 		.extra2		= &maxolduid,
2125 	},
2126 	{
2127 		.procname	= "overflowgid",
2128 		.data		= &overflowgid,
2129 		.maxlen		= sizeof(int),
2130 		.mode		= 0644,
2131 		.proc_handler	= proc_dointvec_minmax,
2132 		.extra1		= &minolduid,
2133 		.extra2		= &maxolduid,
2134 	},
2135 #ifdef CONFIG_S390
2136 	{
2137 		.procname	= "userprocess_debug",
2138 		.data		= &show_unhandled_signals,
2139 		.maxlen		= sizeof(int),
2140 		.mode		= 0644,
2141 		.proc_handler	= proc_dointvec,
2142 	},
2143 #endif
2144 	{
2145 		.procname	= "pid_max",
2146 		.data		= &pid_max,
2147 		.maxlen		= sizeof (int),
2148 		.mode		= 0644,
2149 		.proc_handler	= proc_dointvec_minmax,
2150 		.extra1		= &pid_max_min,
2151 		.extra2		= &pid_max_max,
2152 	},
2153 	{
2154 		.procname	= "panic_on_oops",
2155 		.data		= &panic_on_oops,
2156 		.maxlen		= sizeof(int),
2157 		.mode		= 0644,
2158 		.proc_handler	= proc_dointvec,
2159 	},
2160 	{
2161 		.procname	= "panic_print",
2162 		.data		= &panic_print,
2163 		.maxlen		= sizeof(unsigned long),
2164 		.mode		= 0644,
2165 		.proc_handler	= proc_doulongvec_minmax,
2166 	},
2167 #if defined CONFIG_PRINTK
2168 	{
2169 		.procname	= "printk",
2170 		.data		= &console_loglevel,
2171 		.maxlen		= 4*sizeof(int),
2172 		.mode		= 0644,
2173 		.proc_handler	= proc_dointvec,
2174 	},
2175 	{
2176 		.procname	= "printk_ratelimit",
2177 		.data		= &printk_ratelimit_state.interval,
2178 		.maxlen		= sizeof(int),
2179 		.mode		= 0644,
2180 		.proc_handler	= proc_dointvec_jiffies,
2181 	},
2182 	{
2183 		.procname	= "printk_ratelimit_burst",
2184 		.data		= &printk_ratelimit_state.burst,
2185 		.maxlen		= sizeof(int),
2186 		.mode		= 0644,
2187 		.proc_handler	= proc_dointvec,
2188 	},
2189 	{
2190 		.procname	= "printk_delay",
2191 		.data		= &printk_delay_msec,
2192 		.maxlen		= sizeof(int),
2193 		.mode		= 0644,
2194 		.proc_handler	= proc_dointvec_minmax,
2195 		.extra1		= SYSCTL_ZERO,
2196 		.extra2		= &ten_thousand,
2197 	},
2198 	{
2199 		.procname	= "printk_devkmsg",
2200 		.data		= devkmsg_log_str,
2201 		.maxlen		= DEVKMSG_STR_MAX_SIZE,
2202 		.mode		= 0644,
2203 		.proc_handler	= devkmsg_sysctl_set_loglvl,
2204 	},
2205 	{
2206 		.procname	= "dmesg_restrict",
2207 		.data		= &dmesg_restrict,
2208 		.maxlen		= sizeof(int),
2209 		.mode		= 0644,
2210 		.proc_handler	= proc_dointvec_minmax_sysadmin,
2211 		.extra1		= SYSCTL_ZERO,
2212 		.extra2		= SYSCTL_ONE,
2213 	},
2214 	{
2215 		.procname	= "kptr_restrict",
2216 		.data		= &kptr_restrict,
2217 		.maxlen		= sizeof(int),
2218 		.mode		= 0644,
2219 		.proc_handler	= proc_dointvec_minmax_sysadmin,
2220 		.extra1		= SYSCTL_ZERO,
2221 		.extra2		= &two,
2222 	},
2223 #endif
2224 	{
2225 		.procname	= "ngroups_max",
2226 		.data		= &ngroups_max,
2227 		.maxlen		= sizeof (int),
2228 		.mode		= 0444,
2229 		.proc_handler	= proc_dointvec,
2230 	},
2231 	{
2232 		.procname	= "cap_last_cap",
2233 		.data		= (void *)&cap_last_cap,
2234 		.maxlen		= sizeof(int),
2235 		.mode		= 0444,
2236 		.proc_handler	= proc_dointvec,
2237 	},
2238 #if defined(CONFIG_LOCKUP_DETECTOR)
2239 	{
2240 		.procname       = "watchdog",
2241 		.data		= &watchdog_user_enabled,
2242 		.maxlen		= sizeof(int),
2243 		.mode		= 0644,
2244 		.proc_handler   = proc_watchdog,
2245 		.extra1		= SYSCTL_ZERO,
2246 		.extra2		= SYSCTL_ONE,
2247 	},
2248 	{
2249 		.procname	= "watchdog_thresh",
2250 		.data		= &watchdog_thresh,
2251 		.maxlen		= sizeof(int),
2252 		.mode		= 0644,
2253 		.proc_handler	= proc_watchdog_thresh,
2254 		.extra1		= SYSCTL_ZERO,
2255 		.extra2		= &sixty,
2256 	},
2257 	{
2258 		.procname       = "nmi_watchdog",
2259 		.data		= &nmi_watchdog_user_enabled,
2260 		.maxlen		= sizeof(int),
2261 		.mode		= NMI_WATCHDOG_SYSCTL_PERM,
2262 		.proc_handler   = proc_nmi_watchdog,
2263 		.extra1		= SYSCTL_ZERO,
2264 		.extra2		= SYSCTL_ONE,
2265 	},
2266 	{
2267 		.procname	= "watchdog_cpumask",
2268 		.data		= &watchdog_cpumask_bits,
2269 		.maxlen		= NR_CPUS,
2270 		.mode		= 0644,
2271 		.proc_handler	= proc_watchdog_cpumask,
2272 	},
2273 #ifdef CONFIG_SOFTLOCKUP_DETECTOR
2274 	{
2275 		.procname       = "soft_watchdog",
2276 		.data		= &soft_watchdog_user_enabled,
2277 		.maxlen		= sizeof(int),
2278 		.mode		= 0644,
2279 		.proc_handler   = proc_soft_watchdog,
2280 		.extra1		= SYSCTL_ZERO,
2281 		.extra2		= SYSCTL_ONE,
2282 	},
2283 	{
2284 		.procname	= "softlockup_panic",
2285 		.data		= &softlockup_panic,
2286 		.maxlen		= sizeof(int),
2287 		.mode		= 0644,
2288 		.proc_handler	= proc_dointvec_minmax,
2289 		.extra1		= SYSCTL_ZERO,
2290 		.extra2		= SYSCTL_ONE,
2291 	},
2292 #ifdef CONFIG_SMP
2293 	{
2294 		.procname	= "softlockup_all_cpu_backtrace",
2295 		.data		= &sysctl_softlockup_all_cpu_backtrace,
2296 		.maxlen		= sizeof(int),
2297 		.mode		= 0644,
2298 		.proc_handler	= proc_dointvec_minmax,
2299 		.extra1		= SYSCTL_ZERO,
2300 		.extra2		= SYSCTL_ONE,
2301 	},
2302 #endif /* CONFIG_SMP */
2303 #endif
2304 #ifdef CONFIG_HARDLOCKUP_DETECTOR
2305 	{
2306 		.procname	= "hardlockup_panic",
2307 		.data		= &hardlockup_panic,
2308 		.maxlen		= sizeof(int),
2309 		.mode		= 0644,
2310 		.proc_handler	= proc_dointvec_minmax,
2311 		.extra1		= SYSCTL_ZERO,
2312 		.extra2		= SYSCTL_ONE,
2313 	},
2314 #ifdef CONFIG_SMP
2315 	{
2316 		.procname	= "hardlockup_all_cpu_backtrace",
2317 		.data		= &sysctl_hardlockup_all_cpu_backtrace,
2318 		.maxlen		= sizeof(int),
2319 		.mode		= 0644,
2320 		.proc_handler	= proc_dointvec_minmax,
2321 		.extra1		= SYSCTL_ZERO,
2322 		.extra2		= SYSCTL_ONE,
2323 	},
2324 #endif /* CONFIG_SMP */
2325 #endif
2326 #endif
2327 
2328 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
2329 	{
2330 		.procname       = "unknown_nmi_panic",
2331 		.data           = &unknown_nmi_panic,
2332 		.maxlen         = sizeof (int),
2333 		.mode           = 0644,
2334 		.proc_handler   = proc_dointvec,
2335 	},
2336 #endif
2337 
2338 #if (defined(CONFIG_X86_32) || defined(CONFIG_PARISC)) && \
2339 	defined(CONFIG_DEBUG_STACKOVERFLOW)
2340 	{
2341 		.procname	= "panic_on_stackoverflow",
2342 		.data		= &sysctl_panic_on_stackoverflow,
2343 		.maxlen		= sizeof(int),
2344 		.mode		= 0644,
2345 		.proc_handler	= proc_dointvec,
2346 	},
2347 #endif
2348 #if defined(CONFIG_X86)
2349 	{
2350 		.procname	= "panic_on_unrecovered_nmi",
2351 		.data		= &panic_on_unrecovered_nmi,
2352 		.maxlen		= sizeof(int),
2353 		.mode		= 0644,
2354 		.proc_handler	= proc_dointvec,
2355 	},
2356 	{
2357 		.procname	= "panic_on_io_nmi",
2358 		.data		= &panic_on_io_nmi,
2359 		.maxlen		= sizeof(int),
2360 		.mode		= 0644,
2361 		.proc_handler	= proc_dointvec,
2362 	},
2363 	{
2364 		.procname	= "bootloader_type",
2365 		.data		= &bootloader_type,
2366 		.maxlen		= sizeof (int),
2367 		.mode		= 0444,
2368 		.proc_handler	= proc_dointvec,
2369 	},
2370 	{
2371 		.procname	= "bootloader_version",
2372 		.data		= &bootloader_version,
2373 		.maxlen		= sizeof (int),
2374 		.mode		= 0444,
2375 		.proc_handler	= proc_dointvec,
2376 	},
2377 	{
2378 		.procname	= "io_delay_type",
2379 		.data		= &io_delay_type,
2380 		.maxlen		= sizeof(int),
2381 		.mode		= 0644,
2382 		.proc_handler	= proc_dointvec,
2383 	},
2384 #endif
2385 #if defined(CONFIG_MMU)
2386 	{
2387 		.procname	= "randomize_va_space",
2388 		.data		= &randomize_va_space,
2389 		.maxlen		= sizeof(int),
2390 		.mode		= 0644,
2391 		.proc_handler	= proc_dointvec,
2392 	},
2393 #endif
2394 #if defined(CONFIG_S390) && defined(CONFIG_SMP)
2395 	{
2396 		.procname	= "spin_retry",
2397 		.data		= &spin_retry,
2398 		.maxlen		= sizeof (int),
2399 		.mode		= 0644,
2400 		.proc_handler	= proc_dointvec,
2401 	},
2402 #endif
2403 #if	defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
2404 	{
2405 		.procname	= "acpi_video_flags",
2406 		.data		= &acpi_realmode_flags,
2407 		.maxlen		= sizeof (unsigned long),
2408 		.mode		= 0644,
2409 		.proc_handler	= proc_doulongvec_minmax,
2410 	},
2411 #endif
2412 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
2413 	{
2414 		.procname	= "ignore-unaligned-usertrap",
2415 		.data		= &no_unaligned_warning,
2416 		.maxlen		= sizeof (int),
2417 		.mode		= 0644,
2418 		.proc_handler	= proc_dointvec,
2419 	},
2420 #endif
2421 #ifdef CONFIG_IA64
2422 	{
2423 		.procname	= "unaligned-dump-stack",
2424 		.data		= &unaligned_dump_stack,
2425 		.maxlen		= sizeof (int),
2426 		.mode		= 0644,
2427 		.proc_handler	= proc_dointvec,
2428 	},
2429 #endif
2430 #ifdef CONFIG_DETECT_HUNG_TASK
2431 	{
2432 		.procname	= "hung_task_panic",
2433 		.data		= &sysctl_hung_task_panic,
2434 		.maxlen		= sizeof(int),
2435 		.mode		= 0644,
2436 		.proc_handler	= proc_dointvec_minmax,
2437 		.extra1		= SYSCTL_ZERO,
2438 		.extra2		= SYSCTL_ONE,
2439 	},
2440 	{
2441 		.procname	= "hung_task_check_count",
2442 		.data		= &sysctl_hung_task_check_count,
2443 		.maxlen		= sizeof(int),
2444 		.mode		= 0644,
2445 		.proc_handler	= proc_dointvec_minmax,
2446 		.extra1		= SYSCTL_ZERO,
2447 	},
2448 	{
2449 		.procname	= "hung_task_timeout_secs",
2450 		.data		= &sysctl_hung_task_timeout_secs,
2451 		.maxlen		= sizeof(unsigned long),
2452 		.mode		= 0644,
2453 		.proc_handler	= proc_dohung_task_timeout_secs,
2454 		.extra2		= &hung_task_timeout_max,
2455 	},
2456 	{
2457 		.procname	= "hung_task_check_interval_secs",
2458 		.data		= &sysctl_hung_task_check_interval_secs,
2459 		.maxlen		= sizeof(unsigned long),
2460 		.mode		= 0644,
2461 		.proc_handler	= proc_dohung_task_timeout_secs,
2462 		.extra2		= &hung_task_timeout_max,
2463 	},
2464 	{
2465 		.procname	= "hung_task_warnings",
2466 		.data		= &sysctl_hung_task_warnings,
2467 		.maxlen		= sizeof(int),
2468 		.mode		= 0644,
2469 		.proc_handler	= proc_dointvec_minmax,
2470 		.extra1		= &neg_one,
2471 	},
2472 #endif
2473 #ifdef CONFIG_RT_MUTEXES
2474 	{
2475 		.procname	= "max_lock_depth",
2476 		.data		= &max_lock_depth,
2477 		.maxlen		= sizeof(int),
2478 		.mode		= 0644,
2479 		.proc_handler	= proc_dointvec,
2480 	},
2481 #endif
2482 	{
2483 		.procname	= "poweroff_cmd",
2484 		.data		= &poweroff_cmd,
2485 		.maxlen		= POWEROFF_CMD_PATH_LEN,
2486 		.mode		= 0644,
2487 		.proc_handler	= proc_dostring,
2488 	},
2489 #ifdef CONFIG_KEYS
2490 	{
2491 		.procname	= "keys",
2492 		.mode		= 0555,
2493 		.child		= key_sysctls,
2494 	},
2495 #endif
2496 #ifdef CONFIG_PERF_EVENTS
2497 	/*
2498 	 * User-space scripts rely on the existence of this file
2499 	 * as a feature check for perf_events being enabled.
2500 	 *
2501 	 * So it's an ABI, do not remove!
2502 	 */
2503 	{
2504 		.procname	= "perf_event_paranoid",
2505 		.data		= &sysctl_perf_event_paranoid,
2506 		.maxlen		= sizeof(sysctl_perf_event_paranoid),
2507 		.mode		= 0644,
2508 		.proc_handler	= proc_dointvec,
2509 	},
2510 	{
2511 		.procname	= "perf_event_mlock_kb",
2512 		.data		= &sysctl_perf_event_mlock,
2513 		.maxlen		= sizeof(sysctl_perf_event_mlock),
2514 		.mode		= 0644,
2515 		.proc_handler	= proc_dointvec,
2516 	},
2517 	{
2518 		.procname	= "perf_event_max_sample_rate",
2519 		.data		= &sysctl_perf_event_sample_rate,
2520 		.maxlen		= sizeof(sysctl_perf_event_sample_rate),
2521 		.mode		= 0644,
2522 		.proc_handler	= perf_proc_update_handler,
2523 		.extra1		= SYSCTL_ONE,
2524 	},
2525 	{
2526 		.procname	= "perf_cpu_time_max_percent",
2527 		.data		= &sysctl_perf_cpu_time_max_percent,
2528 		.maxlen		= sizeof(sysctl_perf_cpu_time_max_percent),
2529 		.mode		= 0644,
2530 		.proc_handler	= perf_cpu_time_max_percent_handler,
2531 		.extra1		= SYSCTL_ZERO,
2532 		.extra2		= &one_hundred,
2533 	},
2534 	{
2535 		.procname	= "perf_event_max_stack",
2536 		.data		= &sysctl_perf_event_max_stack,
2537 		.maxlen		= sizeof(sysctl_perf_event_max_stack),
2538 		.mode		= 0644,
2539 		.proc_handler	= perf_event_max_stack_handler,
2540 		.extra1		= SYSCTL_ZERO,
2541 		.extra2		= &six_hundred_forty_kb,
2542 	},
2543 	{
2544 		.procname	= "perf_event_max_contexts_per_stack",
2545 		.data		= &sysctl_perf_event_max_contexts_per_stack,
2546 		.maxlen		= sizeof(sysctl_perf_event_max_contexts_per_stack),
2547 		.mode		= 0644,
2548 		.proc_handler	= perf_event_max_stack_handler,
2549 		.extra1		= SYSCTL_ZERO,
2550 		.extra2		= &one_thousand,
2551 	},
2552 #endif
2553 	{
2554 		.procname	= "panic_on_warn",
2555 		.data		= &panic_on_warn,
2556 		.maxlen		= sizeof(int),
2557 		.mode		= 0644,
2558 		.proc_handler	= proc_dointvec_minmax,
2559 		.extra1		= SYSCTL_ZERO,
2560 		.extra2		= SYSCTL_ONE,
2561 	},
2562 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
2563 	{
2564 		.procname	= "timer_migration",
2565 		.data		= &sysctl_timer_migration,
2566 		.maxlen		= sizeof(unsigned int),
2567 		.mode		= 0644,
2568 		.proc_handler	= timer_migration_handler,
2569 		.extra1		= SYSCTL_ZERO,
2570 		.extra2		= SYSCTL_ONE,
2571 	},
2572 #endif
2573 #ifdef CONFIG_BPF_SYSCALL
2574 	{
2575 		.procname	= "unprivileged_bpf_disabled",
2576 		.data		= &sysctl_unprivileged_bpf_disabled,
2577 		.maxlen		= sizeof(sysctl_unprivileged_bpf_disabled),
2578 		.mode		= 0644,
2579 		/* only handle a transition from default "0" to "1" */
2580 		.proc_handler	= proc_dointvec_minmax,
2581 		.extra1		= SYSCTL_ONE,
2582 		.extra2		= SYSCTL_ONE,
2583 	},
2584 	{
2585 		.procname	= "bpf_stats_enabled",
2586 		.data		= &bpf_stats_enabled_key.key,
2587 		.maxlen		= sizeof(bpf_stats_enabled_key),
2588 		.mode		= 0644,
2589 		.proc_handler	= bpf_stats_handler,
2590 	},
2591 #endif
2592 #if defined(CONFIG_TREE_RCU)
2593 	{
2594 		.procname	= "panic_on_rcu_stall",
2595 		.data		= &sysctl_panic_on_rcu_stall,
2596 		.maxlen		= sizeof(sysctl_panic_on_rcu_stall),
2597 		.mode		= 0644,
2598 		.proc_handler	= proc_dointvec_minmax,
2599 		.extra1		= SYSCTL_ZERO,
2600 		.extra2		= SYSCTL_ONE,
2601 	},
2602 #endif
2603 #ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
2604 	{
2605 		.procname	= "stack_erasing",
2606 		.data		= NULL,
2607 		.maxlen		= sizeof(int),
2608 		.mode		= 0600,
2609 		.proc_handler	= stack_erasing_sysctl,
2610 		.extra1		= SYSCTL_ZERO,
2611 		.extra2		= SYSCTL_ONE,
2612 	},
2613 #endif
2614 	{ }
2615 };
2616 
2617 static struct ctl_table vm_table[] = {
2618 	{
2619 		.procname	= "overcommit_memory",
2620 		.data		= &sysctl_overcommit_memory,
2621 		.maxlen		= sizeof(sysctl_overcommit_memory),
2622 		.mode		= 0644,
2623 		.proc_handler	= proc_dointvec_minmax,
2624 		.extra1		= SYSCTL_ZERO,
2625 		.extra2		= &two,
2626 	},
2627 	{
2628 		.procname	= "panic_on_oom",
2629 		.data		= &sysctl_panic_on_oom,
2630 		.maxlen		= sizeof(sysctl_panic_on_oom),
2631 		.mode		= 0644,
2632 		.proc_handler	= proc_dointvec_minmax,
2633 		.extra1		= SYSCTL_ZERO,
2634 		.extra2		= &two,
2635 	},
2636 	{
2637 		.procname	= "oom_kill_allocating_task",
2638 		.data		= &sysctl_oom_kill_allocating_task,
2639 		.maxlen		= sizeof(sysctl_oom_kill_allocating_task),
2640 		.mode		= 0644,
2641 		.proc_handler	= proc_dointvec,
2642 	},
2643 	{
2644 		.procname	= "oom_dump_tasks",
2645 		.data		= &sysctl_oom_dump_tasks,
2646 		.maxlen		= sizeof(sysctl_oom_dump_tasks),
2647 		.mode		= 0644,
2648 		.proc_handler	= proc_dointvec,
2649 	},
2650 	{
2651 		.procname	= "overcommit_ratio",
2652 		.data		= &sysctl_overcommit_ratio,
2653 		.maxlen		= sizeof(sysctl_overcommit_ratio),
2654 		.mode		= 0644,
2655 		.proc_handler	= overcommit_ratio_handler,
2656 	},
2657 	{
2658 		.procname	= "overcommit_kbytes",
2659 		.data		= &sysctl_overcommit_kbytes,
2660 		.maxlen		= sizeof(sysctl_overcommit_kbytes),
2661 		.mode		= 0644,
2662 		.proc_handler	= overcommit_kbytes_handler,
2663 	},
2664 	{
2665 		.procname	= "page-cluster",
2666 		.data		= &page_cluster,
2667 		.maxlen		= sizeof(int),
2668 		.mode		= 0644,
2669 		.proc_handler	= proc_dointvec_minmax,
2670 		.extra1		= SYSCTL_ZERO,
2671 	},
2672 	{
2673 		.procname	= "dirty_background_ratio",
2674 		.data		= &dirty_background_ratio,
2675 		.maxlen		= sizeof(dirty_background_ratio),
2676 		.mode		= 0644,
2677 		.proc_handler	= dirty_background_ratio_handler,
2678 		.extra1		= SYSCTL_ZERO,
2679 		.extra2		= &one_hundred,
2680 	},
2681 	{
2682 		.procname	= "dirty_background_bytes",
2683 		.data		= &dirty_background_bytes,
2684 		.maxlen		= sizeof(dirty_background_bytes),
2685 		.mode		= 0644,
2686 		.proc_handler	= dirty_background_bytes_handler,
2687 		.extra1		= &one_ul,
2688 	},
2689 	{
2690 		.procname	= "dirty_ratio",
2691 		.data		= &vm_dirty_ratio,
2692 		.maxlen		= sizeof(vm_dirty_ratio),
2693 		.mode		= 0644,
2694 		.proc_handler	= dirty_ratio_handler,
2695 		.extra1		= SYSCTL_ZERO,
2696 		.extra2		= &one_hundred,
2697 	},
2698 	{
2699 		.procname	= "dirty_bytes",
2700 		.data		= &vm_dirty_bytes,
2701 		.maxlen		= sizeof(vm_dirty_bytes),
2702 		.mode		= 0644,
2703 		.proc_handler	= dirty_bytes_handler,
2704 		.extra1		= &dirty_bytes_min,
2705 	},
2706 	{
2707 		.procname	= "dirty_writeback_centisecs",
2708 		.data		= &dirty_writeback_interval,
2709 		.maxlen		= sizeof(dirty_writeback_interval),
2710 		.mode		= 0644,
2711 		.proc_handler	= dirty_writeback_centisecs_handler,
2712 	},
2713 	{
2714 		.procname	= "dirty_expire_centisecs",
2715 		.data		= &dirty_expire_interval,
2716 		.maxlen		= sizeof(dirty_expire_interval),
2717 		.mode		= 0644,
2718 		.proc_handler	= proc_dointvec_minmax,
2719 		.extra1		= SYSCTL_ZERO,
2720 	},
2721 	{
2722 		.procname	= "dirtytime_expire_seconds",
2723 		.data		= &dirtytime_expire_interval,
2724 		.maxlen		= sizeof(dirtytime_expire_interval),
2725 		.mode		= 0644,
2726 		.proc_handler	= dirtytime_interval_handler,
2727 		.extra1		= SYSCTL_ZERO,
2728 	},
2729 	{
2730 		.procname	= "swappiness",
2731 		.data		= &vm_swappiness,
2732 		.maxlen		= sizeof(vm_swappiness),
2733 		.mode		= 0644,
2734 		.proc_handler	= proc_dointvec_minmax,
2735 		.extra1		= SYSCTL_ZERO,
2736 		.extra2		= &two_hundred,
2737 	},
2738 #ifdef CONFIG_HUGETLB_PAGE
2739 	{
2740 		.procname	= "nr_hugepages",
2741 		.data		= NULL,
2742 		.maxlen		= sizeof(unsigned long),
2743 		.mode		= 0644,
2744 		.proc_handler	= hugetlb_sysctl_handler,
2745 	},
2746 #ifdef CONFIG_NUMA
2747 	{
2748 		.procname       = "nr_hugepages_mempolicy",
2749 		.data           = NULL,
2750 		.maxlen         = sizeof(unsigned long),
2751 		.mode           = 0644,
2752 		.proc_handler   = &hugetlb_mempolicy_sysctl_handler,
2753 	},
2754 	{
2755 		.procname		= "numa_stat",
2756 		.data			= &sysctl_vm_numa_stat,
2757 		.maxlen			= sizeof(int),
2758 		.mode			= 0644,
2759 		.proc_handler	= sysctl_vm_numa_stat_handler,
2760 		.extra1			= SYSCTL_ZERO,
2761 		.extra2			= SYSCTL_ONE,
2762 	},
2763 #endif
2764 	 {
2765 		.procname	= "hugetlb_shm_group",
2766 		.data		= &sysctl_hugetlb_shm_group,
2767 		.maxlen		= sizeof(gid_t),
2768 		.mode		= 0644,
2769 		.proc_handler	= proc_dointvec,
2770 	 },
2771 	{
2772 		.procname	= "nr_overcommit_hugepages",
2773 		.data		= NULL,
2774 		.maxlen		= sizeof(unsigned long),
2775 		.mode		= 0644,
2776 		.proc_handler	= hugetlb_overcommit_handler,
2777 	},
2778 #endif
2779 	{
2780 		.procname	= "lowmem_reserve_ratio",
2781 		.data		= &sysctl_lowmem_reserve_ratio,
2782 		.maxlen		= sizeof(sysctl_lowmem_reserve_ratio),
2783 		.mode		= 0644,
2784 		.proc_handler	= lowmem_reserve_ratio_sysctl_handler,
2785 	},
2786 	{
2787 		.procname	= "drop_caches",
2788 		.data		= &sysctl_drop_caches,
2789 		.maxlen		= sizeof(int),
2790 		.mode		= 0200,
2791 		.proc_handler	= drop_caches_sysctl_handler,
2792 		.extra1		= SYSCTL_ONE,
2793 		.extra2		= &four,
2794 	},
2795 #ifdef CONFIG_COMPACTION
2796 	{
2797 		.procname	= "compact_memory",
2798 		.data		= &sysctl_compact_memory,
2799 		.maxlen		= sizeof(int),
2800 		.mode		= 0200,
2801 		.proc_handler	= sysctl_compaction_handler,
2802 	},
2803 	{
2804 		.procname	= "extfrag_threshold",
2805 		.data		= &sysctl_extfrag_threshold,
2806 		.maxlen		= sizeof(int),
2807 		.mode		= 0644,
2808 		.proc_handler	= proc_dointvec_minmax,
2809 		.extra1		= &min_extfrag_threshold,
2810 		.extra2		= &max_extfrag_threshold,
2811 	},
2812 	{
2813 		.procname	= "compact_unevictable_allowed",
2814 		.data		= &sysctl_compact_unevictable_allowed,
2815 		.maxlen		= sizeof(int),
2816 		.mode		= 0644,
2817 		.proc_handler	= proc_dointvec_minmax_warn_RT_change,
2818 		.extra1		= SYSCTL_ZERO,
2819 		.extra2		= SYSCTL_ONE,
2820 	},
2821 
2822 #endif /* CONFIG_COMPACTION */
2823 	{
2824 		.procname	= "min_free_kbytes",
2825 		.data		= &min_free_kbytes,
2826 		.maxlen		= sizeof(min_free_kbytes),
2827 		.mode		= 0644,
2828 		.proc_handler	= min_free_kbytes_sysctl_handler,
2829 		.extra1		= SYSCTL_ZERO,
2830 	},
2831 	{
2832 		.procname	= "watermark_boost_factor",
2833 		.data		= &watermark_boost_factor,
2834 		.maxlen		= sizeof(watermark_boost_factor),
2835 		.mode		= 0644,
2836 		.proc_handler	= proc_dointvec_minmax,
2837 		.extra1		= SYSCTL_ZERO,
2838 	},
2839 	{
2840 		.procname	= "watermark_scale_factor",
2841 		.data		= &watermark_scale_factor,
2842 		.maxlen		= sizeof(watermark_scale_factor),
2843 		.mode		= 0644,
2844 		.proc_handler	= watermark_scale_factor_sysctl_handler,
2845 		.extra1		= SYSCTL_ONE,
2846 		.extra2		= &one_thousand,
2847 	},
2848 	{
2849 		.procname	= "percpu_pagelist_fraction",
2850 		.data		= &percpu_pagelist_fraction,
2851 		.maxlen		= sizeof(percpu_pagelist_fraction),
2852 		.mode		= 0644,
2853 		.proc_handler	= percpu_pagelist_fraction_sysctl_handler,
2854 		.extra1		= SYSCTL_ZERO,
2855 	},
2856 #ifdef CONFIG_MMU
2857 	{
2858 		.procname	= "max_map_count",
2859 		.data		= &sysctl_max_map_count,
2860 		.maxlen		= sizeof(sysctl_max_map_count),
2861 		.mode		= 0644,
2862 		.proc_handler	= proc_dointvec_minmax,
2863 		.extra1		= SYSCTL_ZERO,
2864 	},
2865 #else
2866 	{
2867 		.procname	= "nr_trim_pages",
2868 		.data		= &sysctl_nr_trim_pages,
2869 		.maxlen		= sizeof(sysctl_nr_trim_pages),
2870 		.mode		= 0644,
2871 		.proc_handler	= proc_dointvec_minmax,
2872 		.extra1		= SYSCTL_ZERO,
2873 	},
2874 #endif
2875 	{
2876 		.procname	= "laptop_mode",
2877 		.data		= &laptop_mode,
2878 		.maxlen		= sizeof(laptop_mode),
2879 		.mode		= 0644,
2880 		.proc_handler	= proc_dointvec_jiffies,
2881 	},
2882 	{
2883 		.procname	= "block_dump",
2884 		.data		= &block_dump,
2885 		.maxlen		= sizeof(block_dump),
2886 		.mode		= 0644,
2887 		.proc_handler	= proc_dointvec,
2888 		.extra1		= SYSCTL_ZERO,
2889 	},
2890 	{
2891 		.procname	= "vfs_cache_pressure",
2892 		.data		= &sysctl_vfs_cache_pressure,
2893 		.maxlen		= sizeof(sysctl_vfs_cache_pressure),
2894 		.mode		= 0644,
2895 		.proc_handler	= proc_dointvec,
2896 		.extra1		= SYSCTL_ZERO,
2897 	},
2898 #if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
2899     defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
2900 	{
2901 		.procname	= "legacy_va_layout",
2902 		.data		= &sysctl_legacy_va_layout,
2903 		.maxlen		= sizeof(sysctl_legacy_va_layout),
2904 		.mode		= 0644,
2905 		.proc_handler	= proc_dointvec,
2906 		.extra1		= SYSCTL_ZERO,
2907 	},
2908 #endif
2909 #ifdef CONFIG_NUMA
2910 	{
2911 		.procname	= "zone_reclaim_mode",
2912 		.data		= &node_reclaim_mode,
2913 		.maxlen		= sizeof(node_reclaim_mode),
2914 		.mode		= 0644,
2915 		.proc_handler	= proc_dointvec,
2916 		.extra1		= SYSCTL_ZERO,
2917 	},
2918 	{
2919 		.procname	= "min_unmapped_ratio",
2920 		.data		= &sysctl_min_unmapped_ratio,
2921 		.maxlen		= sizeof(sysctl_min_unmapped_ratio),
2922 		.mode		= 0644,
2923 		.proc_handler	= sysctl_min_unmapped_ratio_sysctl_handler,
2924 		.extra1		= SYSCTL_ZERO,
2925 		.extra2		= &one_hundred,
2926 	},
2927 	{
2928 		.procname	= "min_slab_ratio",
2929 		.data		= &sysctl_min_slab_ratio,
2930 		.maxlen		= sizeof(sysctl_min_slab_ratio),
2931 		.mode		= 0644,
2932 		.proc_handler	= sysctl_min_slab_ratio_sysctl_handler,
2933 		.extra1		= SYSCTL_ZERO,
2934 		.extra2		= &one_hundred,
2935 	},
2936 #endif
2937 #ifdef CONFIG_SMP
2938 	{
2939 		.procname	= "stat_interval",
2940 		.data		= &sysctl_stat_interval,
2941 		.maxlen		= sizeof(sysctl_stat_interval),
2942 		.mode		= 0644,
2943 		.proc_handler	= proc_dointvec_jiffies,
2944 	},
2945 	{
2946 		.procname	= "stat_refresh",
2947 		.data		= NULL,
2948 		.maxlen		= 0,
2949 		.mode		= 0600,
2950 		.proc_handler	= vmstat_refresh,
2951 	},
2952 #endif
2953 #ifdef CONFIG_MMU
2954 	{
2955 		.procname	= "mmap_min_addr",
2956 		.data		= &dac_mmap_min_addr,
2957 		.maxlen		= sizeof(unsigned long),
2958 		.mode		= 0644,
2959 		.proc_handler	= mmap_min_addr_handler,
2960 	},
2961 #endif
2962 #ifdef CONFIG_NUMA
2963 	{
2964 		.procname	= "numa_zonelist_order",
2965 		.data		= &numa_zonelist_order,
2966 		.maxlen		= NUMA_ZONELIST_ORDER_LEN,
2967 		.mode		= 0644,
2968 		.proc_handler	= numa_zonelist_order_handler,
2969 	},
2970 #endif
2971 #if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
2972    (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
2973 	{
2974 		.procname	= "vdso_enabled",
2975 #ifdef CONFIG_X86_32
2976 		.data		= &vdso32_enabled,
2977 		.maxlen		= sizeof(vdso32_enabled),
2978 #else
2979 		.data		= &vdso_enabled,
2980 		.maxlen		= sizeof(vdso_enabled),
2981 #endif
2982 		.mode		= 0644,
2983 		.proc_handler	= proc_dointvec,
2984 		.extra1		= SYSCTL_ZERO,
2985 	},
2986 #endif
2987 #ifdef CONFIG_HIGHMEM
2988 	{
2989 		.procname	= "highmem_is_dirtyable",
2990 		.data		= &vm_highmem_is_dirtyable,
2991 		.maxlen		= sizeof(vm_highmem_is_dirtyable),
2992 		.mode		= 0644,
2993 		.proc_handler	= proc_dointvec_minmax,
2994 		.extra1		= SYSCTL_ZERO,
2995 		.extra2		= SYSCTL_ONE,
2996 	},
2997 #endif
2998 #ifdef CONFIG_MEMORY_FAILURE
2999 	{
3000 		.procname	= "memory_failure_early_kill",
3001 		.data		= &sysctl_memory_failure_early_kill,
3002 		.maxlen		= sizeof(sysctl_memory_failure_early_kill),
3003 		.mode		= 0644,
3004 		.proc_handler	= proc_dointvec_minmax,
3005 		.extra1		= SYSCTL_ZERO,
3006 		.extra2		= SYSCTL_ONE,
3007 	},
3008 	{
3009 		.procname	= "memory_failure_recovery",
3010 		.data		= &sysctl_memory_failure_recovery,
3011 		.maxlen		= sizeof(sysctl_memory_failure_recovery),
3012 		.mode		= 0644,
3013 		.proc_handler	= proc_dointvec_minmax,
3014 		.extra1		= SYSCTL_ZERO,
3015 		.extra2		= SYSCTL_ONE,
3016 	},
3017 #endif
3018 	{
3019 		.procname	= "user_reserve_kbytes",
3020 		.data		= &sysctl_user_reserve_kbytes,
3021 		.maxlen		= sizeof(sysctl_user_reserve_kbytes),
3022 		.mode		= 0644,
3023 		.proc_handler	= proc_doulongvec_minmax,
3024 	},
3025 	{
3026 		.procname	= "admin_reserve_kbytes",
3027 		.data		= &sysctl_admin_reserve_kbytes,
3028 		.maxlen		= sizeof(sysctl_admin_reserve_kbytes),
3029 		.mode		= 0644,
3030 		.proc_handler	= proc_doulongvec_minmax,
3031 	},
3032 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_BITS
3033 	{
3034 		.procname	= "mmap_rnd_bits",
3035 		.data		= &mmap_rnd_bits,
3036 		.maxlen		= sizeof(mmap_rnd_bits),
3037 		.mode		= 0600,
3038 		.proc_handler	= proc_dointvec_minmax,
3039 		.extra1		= (void *)&mmap_rnd_bits_min,
3040 		.extra2		= (void *)&mmap_rnd_bits_max,
3041 	},
3042 #endif
3043 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_COMPAT_BITS
3044 	{
3045 		.procname	= "mmap_rnd_compat_bits",
3046 		.data		= &mmap_rnd_compat_bits,
3047 		.maxlen		= sizeof(mmap_rnd_compat_bits),
3048 		.mode		= 0600,
3049 		.proc_handler	= proc_dointvec_minmax,
3050 		.extra1		= (void *)&mmap_rnd_compat_bits_min,
3051 		.extra2		= (void *)&mmap_rnd_compat_bits_max,
3052 	},
3053 #endif
3054 #ifdef CONFIG_USERFAULTFD
3055 	{
3056 		.procname	= "unprivileged_userfaultfd",
3057 		.data		= &sysctl_unprivileged_userfaultfd,
3058 		.maxlen		= sizeof(sysctl_unprivileged_userfaultfd),
3059 		.mode		= 0644,
3060 		.proc_handler	= proc_dointvec_minmax,
3061 		.extra1		= SYSCTL_ZERO,
3062 		.extra2		= SYSCTL_ONE,
3063 	},
3064 #endif
3065 	{ }
3066 };
3067 
3068 static struct ctl_table fs_table[] = {
3069 	{
3070 		.procname	= "inode-nr",
3071 		.data		= &inodes_stat,
3072 		.maxlen		= 2*sizeof(long),
3073 		.mode		= 0444,
3074 		.proc_handler	= proc_nr_inodes,
3075 	},
3076 	{
3077 		.procname	= "inode-state",
3078 		.data		= &inodes_stat,
3079 		.maxlen		= 7*sizeof(long),
3080 		.mode		= 0444,
3081 		.proc_handler	= proc_nr_inodes,
3082 	},
3083 	{
3084 		.procname	= "file-nr",
3085 		.data		= &files_stat,
3086 		.maxlen		= sizeof(files_stat),
3087 		.mode		= 0444,
3088 		.proc_handler	= proc_nr_files,
3089 	},
3090 	{
3091 		.procname	= "file-max",
3092 		.data		= &files_stat.max_files,
3093 		.maxlen		= sizeof(files_stat.max_files),
3094 		.mode		= 0644,
3095 		.proc_handler	= proc_doulongvec_minmax,
3096 		.extra1		= &zero_ul,
3097 		.extra2		= &long_max,
3098 	},
3099 	{
3100 		.procname	= "nr_open",
3101 		.data		= &sysctl_nr_open,
3102 		.maxlen		= sizeof(unsigned int),
3103 		.mode		= 0644,
3104 		.proc_handler	= proc_dointvec_minmax,
3105 		.extra1		= &sysctl_nr_open_min,
3106 		.extra2		= &sysctl_nr_open_max,
3107 	},
3108 	{
3109 		.procname	= "dentry-state",
3110 		.data		= &dentry_stat,
3111 		.maxlen		= 6*sizeof(long),
3112 		.mode		= 0444,
3113 		.proc_handler	= proc_nr_dentry,
3114 	},
3115 	{
3116 		.procname	= "overflowuid",
3117 		.data		= &fs_overflowuid,
3118 		.maxlen		= sizeof(int),
3119 		.mode		= 0644,
3120 		.proc_handler	= proc_dointvec_minmax,
3121 		.extra1		= &minolduid,
3122 		.extra2		= &maxolduid,
3123 	},
3124 	{
3125 		.procname	= "overflowgid",
3126 		.data		= &fs_overflowgid,
3127 		.maxlen		= sizeof(int),
3128 		.mode		= 0644,
3129 		.proc_handler	= proc_dointvec_minmax,
3130 		.extra1		= &minolduid,
3131 		.extra2		= &maxolduid,
3132 	},
3133 #ifdef CONFIG_FILE_LOCKING
3134 	{
3135 		.procname	= "leases-enable",
3136 		.data		= &leases_enable,
3137 		.maxlen		= sizeof(int),
3138 		.mode		= 0644,
3139 		.proc_handler	= proc_dointvec,
3140 	},
3141 #endif
3142 #ifdef CONFIG_DNOTIFY
3143 	{
3144 		.procname	= "dir-notify-enable",
3145 		.data		= &dir_notify_enable,
3146 		.maxlen		= sizeof(int),
3147 		.mode		= 0644,
3148 		.proc_handler	= proc_dointvec,
3149 	},
3150 #endif
3151 #ifdef CONFIG_MMU
3152 #ifdef CONFIG_FILE_LOCKING
3153 	{
3154 		.procname	= "lease-break-time",
3155 		.data		= &lease_break_time,
3156 		.maxlen		= sizeof(int),
3157 		.mode		= 0644,
3158 		.proc_handler	= proc_dointvec,
3159 	},
3160 #endif
3161 #ifdef CONFIG_AIO
3162 	{
3163 		.procname	= "aio-nr",
3164 		.data		= &aio_nr,
3165 		.maxlen		= sizeof(aio_nr),
3166 		.mode		= 0444,
3167 		.proc_handler	= proc_doulongvec_minmax,
3168 	},
3169 	{
3170 		.procname	= "aio-max-nr",
3171 		.data		= &aio_max_nr,
3172 		.maxlen		= sizeof(aio_max_nr),
3173 		.mode		= 0644,
3174 		.proc_handler	= proc_doulongvec_minmax,
3175 	},
3176 #endif /* CONFIG_AIO */
3177 #ifdef CONFIG_INOTIFY_USER
3178 	{
3179 		.procname	= "inotify",
3180 		.mode		= 0555,
3181 		.child		= inotify_table,
3182 	},
3183 #endif
3184 #ifdef CONFIG_EPOLL
3185 	{
3186 		.procname	= "epoll",
3187 		.mode		= 0555,
3188 		.child		= epoll_table,
3189 	},
3190 #endif
3191 #endif
3192 	{
3193 		.procname	= "protected_symlinks",
3194 		.data		= &sysctl_protected_symlinks,
3195 		.maxlen		= sizeof(int),
3196 		.mode		= 0600,
3197 		.proc_handler	= proc_dointvec_minmax,
3198 		.extra1		= SYSCTL_ZERO,
3199 		.extra2		= SYSCTL_ONE,
3200 	},
3201 	{
3202 		.procname	= "protected_hardlinks",
3203 		.data		= &sysctl_protected_hardlinks,
3204 		.maxlen		= sizeof(int),
3205 		.mode		= 0600,
3206 		.proc_handler	= proc_dointvec_minmax,
3207 		.extra1		= SYSCTL_ZERO,
3208 		.extra2		= SYSCTL_ONE,
3209 	},
3210 	{
3211 		.procname	= "protected_fifos",
3212 		.data		= &sysctl_protected_fifos,
3213 		.maxlen		= sizeof(int),
3214 		.mode		= 0600,
3215 		.proc_handler	= proc_dointvec_minmax,
3216 		.extra1		= SYSCTL_ZERO,
3217 		.extra2		= &two,
3218 	},
3219 	{
3220 		.procname	= "protected_regular",
3221 		.data		= &sysctl_protected_regular,
3222 		.maxlen		= sizeof(int),
3223 		.mode		= 0600,
3224 		.proc_handler	= proc_dointvec_minmax,
3225 		.extra1		= SYSCTL_ZERO,
3226 		.extra2		= &two,
3227 	},
3228 	{
3229 		.procname	= "suid_dumpable",
3230 		.data		= &suid_dumpable,
3231 		.maxlen		= sizeof(int),
3232 		.mode		= 0644,
3233 		.proc_handler	= proc_dointvec_minmax_coredump,
3234 		.extra1		= SYSCTL_ZERO,
3235 		.extra2		= &two,
3236 	},
3237 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
3238 	{
3239 		.procname	= "binfmt_misc",
3240 		.mode		= 0555,
3241 		.child		= sysctl_mount_point,
3242 	},
3243 #endif
3244 	{
3245 		.procname	= "pipe-max-size",
3246 		.data		= &pipe_max_size,
3247 		.maxlen		= sizeof(pipe_max_size),
3248 		.mode		= 0644,
3249 		.proc_handler	= proc_dopipe_max_size,
3250 	},
3251 	{
3252 		.procname	= "pipe-user-pages-hard",
3253 		.data		= &pipe_user_pages_hard,
3254 		.maxlen		= sizeof(pipe_user_pages_hard),
3255 		.mode		= 0644,
3256 		.proc_handler	= proc_doulongvec_minmax,
3257 	},
3258 	{
3259 		.procname	= "pipe-user-pages-soft",
3260 		.data		= &pipe_user_pages_soft,
3261 		.maxlen		= sizeof(pipe_user_pages_soft),
3262 		.mode		= 0644,
3263 		.proc_handler	= proc_doulongvec_minmax,
3264 	},
3265 	{
3266 		.procname	= "mount-max",
3267 		.data		= &sysctl_mount_max,
3268 		.maxlen		= sizeof(unsigned int),
3269 		.mode		= 0644,
3270 		.proc_handler	= proc_dointvec_minmax,
3271 		.extra1		= SYSCTL_ONE,
3272 	},
3273 	{ }
3274 };
3275 
3276 static struct ctl_table debug_table[] = {
3277 #ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
3278 	{
3279 		.procname	= "exception-trace",
3280 		.data		= &show_unhandled_signals,
3281 		.maxlen		= sizeof(int),
3282 		.mode		= 0644,
3283 		.proc_handler	= proc_dointvec
3284 	},
3285 #endif
3286 #if defined(CONFIG_OPTPROBES)
3287 	{
3288 		.procname	= "kprobes-optimization",
3289 		.data		= &sysctl_kprobes_optimization,
3290 		.maxlen		= sizeof(int),
3291 		.mode		= 0644,
3292 		.proc_handler	= proc_kprobes_optimization_handler,
3293 		.extra1		= SYSCTL_ZERO,
3294 		.extra2		= SYSCTL_ONE,
3295 	},
3296 #endif
3297 	{ }
3298 };
3299 
3300 static struct ctl_table dev_table[] = {
3301 	{ }
3302 };
3303 
3304 static struct ctl_table sysctl_base_table[] = {
3305 	{
3306 		.procname	= "kernel",
3307 		.mode		= 0555,
3308 		.child		= kern_table,
3309 	},
3310 	{
3311 		.procname	= "vm",
3312 		.mode		= 0555,
3313 		.child		= vm_table,
3314 	},
3315 	{
3316 		.procname	= "fs",
3317 		.mode		= 0555,
3318 		.child		= fs_table,
3319 	},
3320 	{
3321 		.procname	= "debug",
3322 		.mode		= 0555,
3323 		.child		= debug_table,
3324 	},
3325 	{
3326 		.procname	= "dev",
3327 		.mode		= 0555,
3328 		.child		= dev_table,
3329 	},
3330 	{ }
3331 };
3332 
3333 int __init sysctl_init(void)
3334 {
3335 	struct ctl_table_header *hdr;
3336 
3337 	hdr = register_sysctl_table(sysctl_base_table);
3338 	kmemleak_not_leak(hdr);
3339 	return 0;
3340 }
3341 #endif /* CONFIG_SYSCTL */
3342 /*
3343  * No sense putting this after each symbol definition, twice,
3344  * exception granted :-)
3345  */
3346 EXPORT_SYMBOL(proc_dointvec);
3347 EXPORT_SYMBOL(proc_douintvec);
3348 EXPORT_SYMBOL(proc_dointvec_jiffies);
3349 EXPORT_SYMBOL(proc_dointvec_minmax);
3350 EXPORT_SYMBOL_GPL(proc_douintvec_minmax);
3351 EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
3352 EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
3353 EXPORT_SYMBOL(proc_dostring);
3354 EXPORT_SYMBOL(proc_doulongvec_minmax);
3355 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
3356 EXPORT_SYMBOL(proc_do_large_bitmap);
3357