xref: /openbmc/linux/kernel/sysctl.c (revision 8aaaf2f3af2ae212428f4db1af34214225f5cec3)
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/panic.h>
31 #include <linux/printk.h>
32 #include <linux/proc_fs.h>
33 #include <linux/security.h>
34 #include <linux/ctype.h>
35 #include <linux/kmemleak.h>
36 #include <linux/filter.h>
37 #include <linux/fs.h>
38 #include <linux/init.h>
39 #include <linux/kernel.h>
40 #include <linux/kobject.h>
41 #include <linux/net.h>
42 #include <linux/sysrq.h>
43 #include <linux/highuid.h>
44 #include <linux/writeback.h>
45 #include <linux/ratelimit.h>
46 #include <linux/compaction.h>
47 #include <linux/hugetlb.h>
48 #include <linux/initrd.h>
49 #include <linux/key.h>
50 #include <linux/times.h>
51 #include <linux/limits.h>
52 #include <linux/dcache.h>
53 #include <linux/dnotify.h>
54 #include <linux/syscalls.h>
55 #include <linux/vmstat.h>
56 #include <linux/nfs_fs.h>
57 #include <linux/acpi.h>
58 #include <linux/reboot.h>
59 #include <linux/ftrace.h>
60 #include <linux/perf_event.h>
61 #include <linux/kprobes.h>
62 #include <linux/pipe_fs_i.h>
63 #include <linux/oom.h>
64 #include <linux/kmod.h>
65 #include <linux/capability.h>
66 #include <linux/binfmts.h>
67 #include <linux/sched/sysctl.h>
68 #include <linux/sched/coredump.h>
69 #include <linux/kexec.h>
70 #include <linux/bpf.h>
71 #include <linux/mount.h>
72 #include <linux/userfaultfd_k.h>
73 #include <linux/coredump.h>
74 #include <linux/latencytop.h>
75 #include <linux/pid.h>
76 #include <linux/delayacct.h>
77 
78 #include "../lib/kstrtox.h"
79 
80 #include <linux/uaccess.h>
81 #include <asm/processor.h>
82 
83 #ifdef CONFIG_X86
84 #include <asm/nmi.h>
85 #include <asm/stacktrace.h>
86 #include <asm/io.h>
87 #endif
88 #ifdef CONFIG_SPARC
89 #include <asm/setup.h>
90 #endif
91 #ifdef CONFIG_BSD_PROCESS_ACCT
92 #include <linux/acct.h>
93 #endif
94 #ifdef CONFIG_RT_MUTEXES
95 #include <linux/rtmutex.h>
96 #endif
97 #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
98 #include <linux/lockdep.h>
99 #endif
100 #ifdef CONFIG_CHR_DEV_SG
101 #include <scsi/sg.h>
102 #endif
103 #ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
104 #include <linux/stackleak.h>
105 #endif
106 #ifdef CONFIG_LOCKUP_DETECTOR
107 #include <linux/nmi.h>
108 #endif
109 
110 #if defined(CONFIG_SYSCTL)
111 
112 /* Constants used for minimum and  maximum */
113 #ifdef CONFIG_LOCKUP_DETECTOR
114 static int sixty = 60;
115 #endif
116 
117 static int __maybe_unused neg_one = -1;
118 static int __maybe_unused two = 2;
119 static int __maybe_unused four = 4;
120 static unsigned long zero_ul;
121 static unsigned long one_ul = 1;
122 static unsigned long long_max = LONG_MAX;
123 static int one_hundred = 100;
124 static int two_hundred = 200;
125 static int one_thousand = 1000;
126 #ifdef CONFIG_PRINTK
127 static int ten_thousand = 10000;
128 #endif
129 #ifdef CONFIG_PERF_EVENTS
130 static int six_hundred_forty_kb = 640 * 1024;
131 #endif
132 
133 /* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
134 static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
135 
136 /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
137 static int maxolduid = 65535;
138 static int minolduid;
139 
140 static int ngroups_max = NGROUPS_MAX;
141 static const int cap_last_cap = CAP_LAST_CAP;
142 
143 /*
144  * This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
145  * and hung_task_check_interval_secs
146  */
147 #ifdef CONFIG_DETECT_HUNG_TASK
148 static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
149 #endif
150 
151 #ifdef CONFIG_INOTIFY_USER
152 #include <linux/inotify.h>
153 #endif
154 #ifdef CONFIG_FANOTIFY
155 #include <linux/fanotify.h>
156 #endif
157 
158 #ifdef CONFIG_PROC_SYSCTL
159 
160 /**
161  * enum sysctl_writes_mode - supported sysctl write modes
162  *
163  * @SYSCTL_WRITES_LEGACY: each write syscall must fully contain the sysctl value
164  *	to be written, and multiple writes on the same sysctl file descriptor
165  *	will rewrite the sysctl value, regardless of file position. No warning
166  *	is issued when the initial position is not 0.
167  * @SYSCTL_WRITES_WARN: same as above but warn when the initial file position is
168  *	not 0.
169  * @SYSCTL_WRITES_STRICT: writes to numeric sysctl entries must always be at
170  *	file position 0 and the value must be fully contained in the buffer
171  *	sent to the write syscall. If dealing with strings respect the file
172  *	position, but restrict this to the max length of the buffer, anything
173  *	passed the max length will be ignored. Multiple writes will append
174  *	to the buffer.
175  *
176  * These write modes control how current file position affects the behavior of
177  * updating sysctl values through the proc interface on each write.
178  */
179 enum sysctl_writes_mode {
180 	SYSCTL_WRITES_LEGACY		= -1,
181 	SYSCTL_WRITES_WARN		= 0,
182 	SYSCTL_WRITES_STRICT		= 1,
183 };
184 
185 static enum sysctl_writes_mode sysctl_writes_strict = SYSCTL_WRITES_STRICT;
186 #endif /* CONFIG_PROC_SYSCTL */
187 
188 #if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
189     defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
190 int sysctl_legacy_va_layout;
191 #endif
192 
193 #ifdef CONFIG_COMPACTION
194 static int min_extfrag_threshold;
195 static int max_extfrag_threshold = 1000;
196 #endif
197 
198 #endif /* CONFIG_SYSCTL */
199 
200 #if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_SYSCTL)
201 static int bpf_stats_handler(struct ctl_table *table, int write,
202 			     void *buffer, size_t *lenp, loff_t *ppos)
203 {
204 	struct static_key *key = (struct static_key *)table->data;
205 	static int saved_val;
206 	int val, ret;
207 	struct ctl_table tmp = {
208 		.data   = &val,
209 		.maxlen = sizeof(val),
210 		.mode   = table->mode,
211 		.extra1 = SYSCTL_ZERO,
212 		.extra2 = SYSCTL_ONE,
213 	};
214 
215 	if (write && !capable(CAP_SYS_ADMIN))
216 		return -EPERM;
217 
218 	mutex_lock(&bpf_stats_enabled_mutex);
219 	val = saved_val;
220 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
221 	if (write && !ret && val != saved_val) {
222 		if (val)
223 			static_key_slow_inc(key);
224 		else
225 			static_key_slow_dec(key);
226 		saved_val = val;
227 	}
228 	mutex_unlock(&bpf_stats_enabled_mutex);
229 	return ret;
230 }
231 
232 static int bpf_unpriv_handler(struct ctl_table *table, int write,
233 			      void *buffer, size_t *lenp, loff_t *ppos)
234 {
235 	int ret, unpriv_enable = *(int *)table->data;
236 	bool locked_state = unpriv_enable == 1;
237 	struct ctl_table tmp = *table;
238 
239 	if (write && !capable(CAP_SYS_ADMIN))
240 		return -EPERM;
241 
242 	tmp.data = &unpriv_enable;
243 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
244 	if (write && !ret) {
245 		if (locked_state && unpriv_enable != 1)
246 			return -EPERM;
247 		*(int *)table->data = unpriv_enable;
248 	}
249 	return ret;
250 }
251 #endif /* CONFIG_BPF_SYSCALL && CONFIG_SYSCTL */
252 
253 /*
254  * /proc/sys support
255  */
256 
257 #ifdef CONFIG_PROC_SYSCTL
258 
259 static int _proc_do_string(char *data, int maxlen, int write,
260 		char *buffer, size_t *lenp, loff_t *ppos)
261 {
262 	size_t len;
263 	char c, *p;
264 
265 	if (!data || !maxlen || !*lenp) {
266 		*lenp = 0;
267 		return 0;
268 	}
269 
270 	if (write) {
271 		if (sysctl_writes_strict == SYSCTL_WRITES_STRICT) {
272 			/* Only continue writes not past the end of buffer. */
273 			len = strlen(data);
274 			if (len > maxlen - 1)
275 				len = maxlen - 1;
276 
277 			if (*ppos > len)
278 				return 0;
279 			len = *ppos;
280 		} else {
281 			/* Start writing from beginning of buffer. */
282 			len = 0;
283 		}
284 
285 		*ppos += *lenp;
286 		p = buffer;
287 		while ((p - buffer) < *lenp && len < maxlen - 1) {
288 			c = *(p++);
289 			if (c == 0 || c == '\n')
290 				break;
291 			data[len++] = c;
292 		}
293 		data[len] = 0;
294 	} else {
295 		len = strlen(data);
296 		if (len > maxlen)
297 			len = maxlen;
298 
299 		if (*ppos > len) {
300 			*lenp = 0;
301 			return 0;
302 		}
303 
304 		data += *ppos;
305 		len  -= *ppos;
306 
307 		if (len > *lenp)
308 			len = *lenp;
309 		if (len)
310 			memcpy(buffer, data, len);
311 		if (len < *lenp) {
312 			buffer[len] = '\n';
313 			len++;
314 		}
315 		*lenp = len;
316 		*ppos += len;
317 	}
318 	return 0;
319 }
320 
321 static void warn_sysctl_write(struct ctl_table *table)
322 {
323 	pr_warn_once("%s wrote to %s when file position was not 0!\n"
324 		"This will not be supported in the future. To silence this\n"
325 		"warning, set kernel.sysctl_writes_strict = -1\n",
326 		current->comm, table->procname);
327 }
328 
329 /**
330  * proc_first_pos_non_zero_ignore - check if first position is allowed
331  * @ppos: file position
332  * @table: the sysctl table
333  *
334  * Returns true if the first position is non-zero and the sysctl_writes_strict
335  * mode indicates this is not allowed for numeric input types. String proc
336  * handlers can ignore the return value.
337  */
338 static bool proc_first_pos_non_zero_ignore(loff_t *ppos,
339 					   struct ctl_table *table)
340 {
341 	if (!*ppos)
342 		return false;
343 
344 	switch (sysctl_writes_strict) {
345 	case SYSCTL_WRITES_STRICT:
346 		return true;
347 	case SYSCTL_WRITES_WARN:
348 		warn_sysctl_write(table);
349 		return false;
350 	default:
351 		return false;
352 	}
353 }
354 
355 /**
356  * proc_dostring - read a string sysctl
357  * @table: the sysctl table
358  * @write: %TRUE if this is a write to the sysctl file
359  * @buffer: the user buffer
360  * @lenp: the size of the user buffer
361  * @ppos: file position
362  *
363  * Reads/writes a string from/to the user buffer. If the kernel
364  * buffer provided is not large enough to hold the string, the
365  * string is truncated. The copied string is %NULL-terminated.
366  * If the string is being read by the user process, it is copied
367  * and a newline '\n' is added. It is truncated if the buffer is
368  * not large enough.
369  *
370  * Returns 0 on success.
371  */
372 int proc_dostring(struct ctl_table *table, int write,
373 		  void *buffer, size_t *lenp, loff_t *ppos)
374 {
375 	if (write)
376 		proc_first_pos_non_zero_ignore(ppos, table);
377 
378 	return _proc_do_string(table->data, table->maxlen, write, buffer, lenp,
379 			ppos);
380 }
381 
382 static size_t proc_skip_spaces(char **buf)
383 {
384 	size_t ret;
385 	char *tmp = skip_spaces(*buf);
386 	ret = tmp - *buf;
387 	*buf = tmp;
388 	return ret;
389 }
390 
391 static void proc_skip_char(char **buf, size_t *size, const char v)
392 {
393 	while (*size) {
394 		if (**buf != v)
395 			break;
396 		(*size)--;
397 		(*buf)++;
398 	}
399 }
400 
401 /**
402  * strtoul_lenient - parse an ASCII formatted integer from a buffer and only
403  *                   fail on overflow
404  *
405  * @cp: kernel buffer containing the string to parse
406  * @endp: pointer to store the trailing characters
407  * @base: the base to use
408  * @res: where the parsed integer will be stored
409  *
410  * In case of success 0 is returned and @res will contain the parsed integer,
411  * @endp will hold any trailing characters.
412  * This function will fail the parse on overflow. If there wasn't an overflow
413  * the function will defer the decision what characters count as invalid to the
414  * caller.
415  */
416 static int strtoul_lenient(const char *cp, char **endp, unsigned int base,
417 			   unsigned long *res)
418 {
419 	unsigned long long result;
420 	unsigned int rv;
421 
422 	cp = _parse_integer_fixup_radix(cp, &base);
423 	rv = _parse_integer(cp, base, &result);
424 	if ((rv & KSTRTOX_OVERFLOW) || (result != (unsigned long)result))
425 		return -ERANGE;
426 
427 	cp += rv;
428 
429 	if (endp)
430 		*endp = (char *)cp;
431 
432 	*res = (unsigned long)result;
433 	return 0;
434 }
435 
436 #define TMPBUFLEN 22
437 /**
438  * proc_get_long - reads an ASCII formatted integer from a user buffer
439  *
440  * @buf: a kernel buffer
441  * @size: size of the kernel buffer
442  * @val: this is where the number will be stored
443  * @neg: set to %TRUE if number is negative
444  * @perm_tr: a vector which contains the allowed trailers
445  * @perm_tr_len: size of the perm_tr vector
446  * @tr: pointer to store the trailer character
447  *
448  * In case of success %0 is returned and @buf and @size are updated with
449  * the amount of bytes read. If @tr is non-NULL and a trailing
450  * character exists (size is non-zero after returning from this
451  * function), @tr is updated with the trailing character.
452  */
453 static int proc_get_long(char **buf, size_t *size,
454 			  unsigned long *val, bool *neg,
455 			  const char *perm_tr, unsigned perm_tr_len, char *tr)
456 {
457 	int len;
458 	char *p, tmp[TMPBUFLEN];
459 
460 	if (!*size)
461 		return -EINVAL;
462 
463 	len = *size;
464 	if (len > TMPBUFLEN - 1)
465 		len = TMPBUFLEN - 1;
466 
467 	memcpy(tmp, *buf, len);
468 
469 	tmp[len] = 0;
470 	p = tmp;
471 	if (*p == '-' && *size > 1) {
472 		*neg = true;
473 		p++;
474 	} else
475 		*neg = false;
476 	if (!isdigit(*p))
477 		return -EINVAL;
478 
479 	if (strtoul_lenient(p, &p, 0, val))
480 		return -EINVAL;
481 
482 	len = p - tmp;
483 
484 	/* We don't know if the next char is whitespace thus we may accept
485 	 * invalid integers (e.g. 1234...a) or two integers instead of one
486 	 * (e.g. 123...1). So lets not allow such large numbers. */
487 	if (len == TMPBUFLEN - 1)
488 		return -EINVAL;
489 
490 	if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
491 		return -EINVAL;
492 
493 	if (tr && (len < *size))
494 		*tr = *p;
495 
496 	*buf += len;
497 	*size -= len;
498 
499 	return 0;
500 }
501 
502 /**
503  * proc_put_long - converts an integer to a decimal ASCII formatted string
504  *
505  * @buf: the user buffer
506  * @size: the size of the user buffer
507  * @val: the integer to be converted
508  * @neg: sign of the number, %TRUE for negative
509  *
510  * In case of success @buf and @size are updated with the amount of bytes
511  * written.
512  */
513 static void proc_put_long(void **buf, size_t *size, unsigned long val, bool neg)
514 {
515 	int len;
516 	char tmp[TMPBUFLEN], *p = tmp;
517 
518 	sprintf(p, "%s%lu", neg ? "-" : "", val);
519 	len = strlen(tmp);
520 	if (len > *size)
521 		len = *size;
522 	memcpy(*buf, tmp, len);
523 	*size -= len;
524 	*buf += len;
525 }
526 #undef TMPBUFLEN
527 
528 static void proc_put_char(void **buf, size_t *size, char c)
529 {
530 	if (*size) {
531 		char **buffer = (char **)buf;
532 		**buffer = c;
533 
534 		(*size)--;
535 		(*buffer)++;
536 		*buf = *buffer;
537 	}
538 }
539 
540 static int do_proc_dobool_conv(bool *negp, unsigned long *lvalp,
541 				int *valp,
542 				int write, void *data)
543 {
544 	if (write) {
545 		*(bool *)valp = *lvalp;
546 	} else {
547 		int val = *(bool *)valp;
548 
549 		*lvalp = (unsigned long)val;
550 		*negp = false;
551 	}
552 	return 0;
553 }
554 
555 static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
556 				 int *valp,
557 				 int write, void *data)
558 {
559 	if (write) {
560 		if (*negp) {
561 			if (*lvalp > (unsigned long) INT_MAX + 1)
562 				return -EINVAL;
563 			*valp = -*lvalp;
564 		} else {
565 			if (*lvalp > (unsigned long) INT_MAX)
566 				return -EINVAL;
567 			*valp = *lvalp;
568 		}
569 	} else {
570 		int val = *valp;
571 		if (val < 0) {
572 			*negp = true;
573 			*lvalp = -(unsigned long)val;
574 		} else {
575 			*negp = false;
576 			*lvalp = (unsigned long)val;
577 		}
578 	}
579 	return 0;
580 }
581 
582 static int do_proc_douintvec_conv(unsigned long *lvalp,
583 				  unsigned int *valp,
584 				  int write, void *data)
585 {
586 	if (write) {
587 		if (*lvalp > UINT_MAX)
588 			return -EINVAL;
589 		*valp = *lvalp;
590 	} else {
591 		unsigned int val = *valp;
592 		*lvalp = (unsigned long)val;
593 	}
594 	return 0;
595 }
596 
597 static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
598 
599 static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
600 		  int write, void *buffer,
601 		  size_t *lenp, loff_t *ppos,
602 		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
603 			      int write, void *data),
604 		  void *data)
605 {
606 	int *i, vleft, first = 1, err = 0;
607 	size_t left;
608 	char *p;
609 
610 	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
611 		*lenp = 0;
612 		return 0;
613 	}
614 
615 	i = (int *) tbl_data;
616 	vleft = table->maxlen / sizeof(*i);
617 	left = *lenp;
618 
619 	if (!conv)
620 		conv = do_proc_dointvec_conv;
621 
622 	if (write) {
623 		if (proc_first_pos_non_zero_ignore(ppos, table))
624 			goto out;
625 
626 		if (left > PAGE_SIZE - 1)
627 			left = PAGE_SIZE - 1;
628 		p = buffer;
629 	}
630 
631 	for (; left && vleft--; i++, first=0) {
632 		unsigned long lval;
633 		bool neg;
634 
635 		if (write) {
636 			left -= proc_skip_spaces(&p);
637 
638 			if (!left)
639 				break;
640 			err = proc_get_long(&p, &left, &lval, &neg,
641 					     proc_wspace_sep,
642 					     sizeof(proc_wspace_sep), NULL);
643 			if (err)
644 				break;
645 			if (conv(&neg, &lval, i, 1, data)) {
646 				err = -EINVAL;
647 				break;
648 			}
649 		} else {
650 			if (conv(&neg, &lval, i, 0, data)) {
651 				err = -EINVAL;
652 				break;
653 			}
654 			if (!first)
655 				proc_put_char(&buffer, &left, '\t');
656 			proc_put_long(&buffer, &left, lval, neg);
657 		}
658 	}
659 
660 	if (!write && !first && left && !err)
661 		proc_put_char(&buffer, &left, '\n');
662 	if (write && !err && left)
663 		left -= proc_skip_spaces(&p);
664 	if (write && first)
665 		return err ? : -EINVAL;
666 	*lenp -= left;
667 out:
668 	*ppos += *lenp;
669 	return err;
670 }
671 
672 static int do_proc_dointvec(struct ctl_table *table, int write,
673 		  void *buffer, size_t *lenp, loff_t *ppos,
674 		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
675 			      int write, void *data),
676 		  void *data)
677 {
678 	return __do_proc_dointvec(table->data, table, write,
679 			buffer, lenp, ppos, conv, data);
680 }
681 
682 static int do_proc_douintvec_w(unsigned int *tbl_data,
683 			       struct ctl_table *table,
684 			       void *buffer,
685 			       size_t *lenp, loff_t *ppos,
686 			       int (*conv)(unsigned long *lvalp,
687 					   unsigned int *valp,
688 					   int write, void *data),
689 			       void *data)
690 {
691 	unsigned long lval;
692 	int err = 0;
693 	size_t left;
694 	bool neg;
695 	char *p = buffer;
696 
697 	left = *lenp;
698 
699 	if (proc_first_pos_non_zero_ignore(ppos, table))
700 		goto bail_early;
701 
702 	if (left > PAGE_SIZE - 1)
703 		left = PAGE_SIZE - 1;
704 
705 	left -= proc_skip_spaces(&p);
706 	if (!left) {
707 		err = -EINVAL;
708 		goto out_free;
709 	}
710 
711 	err = proc_get_long(&p, &left, &lval, &neg,
712 			     proc_wspace_sep,
713 			     sizeof(proc_wspace_sep), NULL);
714 	if (err || neg) {
715 		err = -EINVAL;
716 		goto out_free;
717 	}
718 
719 	if (conv(&lval, tbl_data, 1, data)) {
720 		err = -EINVAL;
721 		goto out_free;
722 	}
723 
724 	if (!err && left)
725 		left -= proc_skip_spaces(&p);
726 
727 out_free:
728 	if (err)
729 		return -EINVAL;
730 
731 	return 0;
732 
733 	/* This is in keeping with old __do_proc_dointvec() */
734 bail_early:
735 	*ppos += *lenp;
736 	return err;
737 }
738 
739 static int do_proc_douintvec_r(unsigned int *tbl_data, void *buffer,
740 			       size_t *lenp, loff_t *ppos,
741 			       int (*conv)(unsigned long *lvalp,
742 					   unsigned int *valp,
743 					   int write, void *data),
744 			       void *data)
745 {
746 	unsigned long lval;
747 	int err = 0;
748 	size_t left;
749 
750 	left = *lenp;
751 
752 	if (conv(&lval, tbl_data, 0, data)) {
753 		err = -EINVAL;
754 		goto out;
755 	}
756 
757 	proc_put_long(&buffer, &left, lval, false);
758 	if (!left)
759 		goto out;
760 
761 	proc_put_char(&buffer, &left, '\n');
762 
763 out:
764 	*lenp -= left;
765 	*ppos += *lenp;
766 
767 	return err;
768 }
769 
770 static int __do_proc_douintvec(void *tbl_data, struct ctl_table *table,
771 			       int write, void *buffer,
772 			       size_t *lenp, loff_t *ppos,
773 			       int (*conv)(unsigned long *lvalp,
774 					   unsigned int *valp,
775 					   int write, void *data),
776 			       void *data)
777 {
778 	unsigned int *i, vleft;
779 
780 	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
781 		*lenp = 0;
782 		return 0;
783 	}
784 
785 	i = (unsigned int *) tbl_data;
786 	vleft = table->maxlen / sizeof(*i);
787 
788 	/*
789 	 * Arrays are not supported, keep this simple. *Do not* add
790 	 * support for them.
791 	 */
792 	if (vleft != 1) {
793 		*lenp = 0;
794 		return -EINVAL;
795 	}
796 
797 	if (!conv)
798 		conv = do_proc_douintvec_conv;
799 
800 	if (write)
801 		return do_proc_douintvec_w(i, table, buffer, lenp, ppos,
802 					   conv, data);
803 	return do_proc_douintvec_r(i, buffer, lenp, ppos, conv, data);
804 }
805 
806 static int do_proc_douintvec(struct ctl_table *table, int write,
807 			     void *buffer, size_t *lenp, loff_t *ppos,
808 			     int (*conv)(unsigned long *lvalp,
809 					 unsigned int *valp,
810 					 int write, void *data),
811 			     void *data)
812 {
813 	return __do_proc_douintvec(table->data, table, write,
814 				   buffer, lenp, ppos, conv, data);
815 }
816 
817 /**
818  * proc_dobool - read/write a bool
819  * @table: the sysctl table
820  * @write: %TRUE if this is a write to the sysctl file
821  * @buffer: the user buffer
822  * @lenp: the size of the user buffer
823  * @ppos: file position
824  *
825  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
826  * values from/to the user buffer, treated as an ASCII string.
827  *
828  * Returns 0 on success.
829  */
830 int proc_dobool(struct ctl_table *table, int write, void *buffer,
831 		size_t *lenp, loff_t *ppos)
832 {
833 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
834 				do_proc_dobool_conv, NULL);
835 }
836 
837 /**
838  * proc_dointvec - read a vector of integers
839  * @table: the sysctl table
840  * @write: %TRUE if this is a write to the sysctl file
841  * @buffer: the user buffer
842  * @lenp: the size of the user buffer
843  * @ppos: file position
844  *
845  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
846  * values from/to the user buffer, treated as an ASCII string.
847  *
848  * Returns 0 on success.
849  */
850 int proc_dointvec(struct ctl_table *table, int write, void *buffer,
851 		  size_t *lenp, loff_t *ppos)
852 {
853 	return do_proc_dointvec(table, write, buffer, lenp, ppos, NULL, NULL);
854 }
855 
856 #ifdef CONFIG_COMPACTION
857 static int proc_dointvec_minmax_warn_RT_change(struct ctl_table *table,
858 		int write, void *buffer, size_t *lenp, loff_t *ppos)
859 {
860 	int ret, old;
861 
862 	if (!IS_ENABLED(CONFIG_PREEMPT_RT) || !write)
863 		return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
864 
865 	old = *(int *)table->data;
866 	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
867 	if (ret)
868 		return ret;
869 	if (old != *(int *)table->data)
870 		pr_warn_once("sysctl attribute %s changed by %s[%d]\n",
871 			     table->procname, current->comm,
872 			     task_pid_nr(current));
873 	return ret;
874 }
875 #endif
876 
877 /**
878  * proc_douintvec - read a vector of unsigned integers
879  * @table: the sysctl table
880  * @write: %TRUE if this is a write to the sysctl file
881  * @buffer: the user buffer
882  * @lenp: the size of the user buffer
883  * @ppos: file position
884  *
885  * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
886  * values from/to the user buffer, treated as an ASCII string.
887  *
888  * Returns 0 on success.
889  */
890 int proc_douintvec(struct ctl_table *table, int write, void *buffer,
891 		size_t *lenp, loff_t *ppos)
892 {
893 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
894 				 do_proc_douintvec_conv, NULL);
895 }
896 
897 /*
898  * Taint values can only be increased
899  * This means we can safely use a temporary.
900  */
901 static int proc_taint(struct ctl_table *table, int write,
902 			       void *buffer, size_t *lenp, loff_t *ppos)
903 {
904 	struct ctl_table t;
905 	unsigned long tmptaint = get_taint();
906 	int err;
907 
908 	if (write && !capable(CAP_SYS_ADMIN))
909 		return -EPERM;
910 
911 	t = *table;
912 	t.data = &tmptaint;
913 	err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
914 	if (err < 0)
915 		return err;
916 
917 	if (write) {
918 		int i;
919 
920 		/*
921 		 * If we are relying on panic_on_taint not producing
922 		 * false positives due to userspace input, bail out
923 		 * before setting the requested taint flags.
924 		 */
925 		if (panic_on_taint_nousertaint && (tmptaint & panic_on_taint))
926 			return -EINVAL;
927 
928 		/*
929 		 * Poor man's atomic or. Not worth adding a primitive
930 		 * to everyone's atomic.h for this
931 		 */
932 		for (i = 0; i < TAINT_FLAGS_COUNT; i++)
933 			if ((1UL << i) & tmptaint)
934 				add_taint(i, LOCKDEP_STILL_OK);
935 	}
936 
937 	return err;
938 }
939 
940 #ifdef CONFIG_PRINTK
941 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
942 				void *buffer, size_t *lenp, loff_t *ppos)
943 {
944 	if (write && !capable(CAP_SYS_ADMIN))
945 		return -EPERM;
946 
947 	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
948 }
949 #endif
950 
951 /**
952  * struct do_proc_dointvec_minmax_conv_param - proc_dointvec_minmax() range checking structure
953  * @min: pointer to minimum allowable value
954  * @max: pointer to maximum allowable value
955  *
956  * The do_proc_dointvec_minmax_conv_param structure provides the
957  * minimum and maximum values for doing range checking for those sysctl
958  * parameters that use the proc_dointvec_minmax() handler.
959  */
960 struct do_proc_dointvec_minmax_conv_param {
961 	int *min;
962 	int *max;
963 };
964 
965 static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
966 					int *valp,
967 					int write, void *data)
968 {
969 	int tmp, ret;
970 	struct do_proc_dointvec_minmax_conv_param *param = data;
971 	/*
972 	 * If writing, first do so via a temporary local int so we can
973 	 * bounds-check it before touching *valp.
974 	 */
975 	int *ip = write ? &tmp : valp;
976 
977 	ret = do_proc_dointvec_conv(negp, lvalp, ip, write, data);
978 	if (ret)
979 		return ret;
980 
981 	if (write) {
982 		if ((param->min && *param->min > tmp) ||
983 		    (param->max && *param->max < tmp))
984 			return -EINVAL;
985 		*valp = tmp;
986 	}
987 
988 	return 0;
989 }
990 
991 /**
992  * proc_dointvec_minmax - read a vector of integers with min/max values
993  * @table: the sysctl table
994  * @write: %TRUE if this is a write to the sysctl file
995  * @buffer: the user buffer
996  * @lenp: the size of the user buffer
997  * @ppos: file position
998  *
999  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1000  * values from/to the user buffer, treated as an ASCII string.
1001  *
1002  * This routine will ensure the values are within the range specified by
1003  * table->extra1 (min) and table->extra2 (max).
1004  *
1005  * Returns 0 on success or -EINVAL on write when the range check fails.
1006  */
1007 int proc_dointvec_minmax(struct ctl_table *table, int write,
1008 		  void *buffer, size_t *lenp, loff_t *ppos)
1009 {
1010 	struct do_proc_dointvec_minmax_conv_param param = {
1011 		.min = (int *) table->extra1,
1012 		.max = (int *) table->extra2,
1013 	};
1014 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
1015 				do_proc_dointvec_minmax_conv, &param);
1016 }
1017 
1018 /**
1019  * struct do_proc_douintvec_minmax_conv_param - proc_douintvec_minmax() range checking structure
1020  * @min: pointer to minimum allowable value
1021  * @max: pointer to maximum allowable value
1022  *
1023  * The do_proc_douintvec_minmax_conv_param structure provides the
1024  * minimum and maximum values for doing range checking for those sysctl
1025  * parameters that use the proc_douintvec_minmax() handler.
1026  */
1027 struct do_proc_douintvec_minmax_conv_param {
1028 	unsigned int *min;
1029 	unsigned int *max;
1030 };
1031 
1032 static int do_proc_douintvec_minmax_conv(unsigned long *lvalp,
1033 					 unsigned int *valp,
1034 					 int write, void *data)
1035 {
1036 	int ret;
1037 	unsigned int tmp;
1038 	struct do_proc_douintvec_minmax_conv_param *param = data;
1039 	/* write via temporary local uint for bounds-checking */
1040 	unsigned int *up = write ? &tmp : valp;
1041 
1042 	ret = do_proc_douintvec_conv(lvalp, up, write, data);
1043 	if (ret)
1044 		return ret;
1045 
1046 	if (write) {
1047 		if ((param->min && *param->min > tmp) ||
1048 		    (param->max && *param->max < tmp))
1049 			return -ERANGE;
1050 
1051 		*valp = tmp;
1052 	}
1053 
1054 	return 0;
1055 }
1056 
1057 /**
1058  * proc_douintvec_minmax - read a vector of unsigned ints with min/max values
1059  * @table: the sysctl table
1060  * @write: %TRUE if this is a write to the sysctl file
1061  * @buffer: the user buffer
1062  * @lenp: the size of the user buffer
1063  * @ppos: file position
1064  *
1065  * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
1066  * values from/to the user buffer, treated as an ASCII string. Negative
1067  * strings are not allowed.
1068  *
1069  * This routine will ensure the values are within the range specified by
1070  * table->extra1 (min) and table->extra2 (max). There is a final sanity
1071  * check for UINT_MAX to avoid having to support wrap around uses from
1072  * userspace.
1073  *
1074  * Returns 0 on success or -ERANGE on write when the range check fails.
1075  */
1076 int proc_douintvec_minmax(struct ctl_table *table, int write,
1077 			  void *buffer, size_t *lenp, loff_t *ppos)
1078 {
1079 	struct do_proc_douintvec_minmax_conv_param param = {
1080 		.min = (unsigned int *) table->extra1,
1081 		.max = (unsigned int *) table->extra2,
1082 	};
1083 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1084 				 do_proc_douintvec_minmax_conv, &param);
1085 }
1086 
1087 /**
1088  * proc_dou8vec_minmax - read a vector of unsigned chars with min/max values
1089  * @table: the sysctl table
1090  * @write: %TRUE if this is a write to the sysctl file
1091  * @buffer: the user buffer
1092  * @lenp: the size of the user buffer
1093  * @ppos: file position
1094  *
1095  * Reads/writes up to table->maxlen/sizeof(u8) unsigned chars
1096  * values from/to the user buffer, treated as an ASCII string. Negative
1097  * strings are not allowed.
1098  *
1099  * This routine will ensure the values are within the range specified by
1100  * table->extra1 (min) and table->extra2 (max).
1101  *
1102  * Returns 0 on success or an error on write when the range check fails.
1103  */
1104 int proc_dou8vec_minmax(struct ctl_table *table, int write,
1105 			void *buffer, size_t *lenp, loff_t *ppos)
1106 {
1107 	struct ctl_table tmp;
1108 	unsigned int min = 0, max = 255U, val;
1109 	u8 *data = table->data;
1110 	struct do_proc_douintvec_minmax_conv_param param = {
1111 		.min = &min,
1112 		.max = &max,
1113 	};
1114 	int res;
1115 
1116 	/* Do not support arrays yet. */
1117 	if (table->maxlen != sizeof(u8))
1118 		return -EINVAL;
1119 
1120 	if (table->extra1) {
1121 		min = *(unsigned int *) table->extra1;
1122 		if (min > 255U)
1123 			return -EINVAL;
1124 	}
1125 	if (table->extra2) {
1126 		max = *(unsigned int *) table->extra2;
1127 		if (max > 255U)
1128 			return -EINVAL;
1129 	}
1130 
1131 	tmp = *table;
1132 
1133 	tmp.maxlen = sizeof(val);
1134 	tmp.data = &val;
1135 	val = *data;
1136 	res = do_proc_douintvec(&tmp, write, buffer, lenp, ppos,
1137 				do_proc_douintvec_minmax_conv, &param);
1138 	if (res)
1139 		return res;
1140 	if (write)
1141 		*data = val;
1142 	return 0;
1143 }
1144 EXPORT_SYMBOL_GPL(proc_dou8vec_minmax);
1145 
1146 static int do_proc_dopipe_max_size_conv(unsigned long *lvalp,
1147 					unsigned int *valp,
1148 					int write, void *data)
1149 {
1150 	if (write) {
1151 		unsigned int val;
1152 
1153 		val = round_pipe_size(*lvalp);
1154 		if (val == 0)
1155 			return -EINVAL;
1156 
1157 		*valp = val;
1158 	} else {
1159 		unsigned int val = *valp;
1160 		*lvalp = (unsigned long) val;
1161 	}
1162 
1163 	return 0;
1164 }
1165 
1166 static int proc_dopipe_max_size(struct ctl_table *table, int write,
1167 				void *buffer, size_t *lenp, loff_t *ppos)
1168 {
1169 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1170 				 do_proc_dopipe_max_size_conv, NULL);
1171 }
1172 
1173 static void validate_coredump_safety(void)
1174 {
1175 #ifdef CONFIG_COREDUMP
1176 	if (suid_dumpable == SUID_DUMP_ROOT &&
1177 	    core_pattern[0] != '/' && core_pattern[0] != '|') {
1178 		printk(KERN_WARNING
1179 "Unsafe core_pattern used with fs.suid_dumpable=2.\n"
1180 "Pipe handler or fully qualified core dump path required.\n"
1181 "Set kernel.core_pattern before fs.suid_dumpable.\n"
1182 		);
1183 	}
1184 #endif
1185 }
1186 
1187 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
1188 		void *buffer, size_t *lenp, loff_t *ppos)
1189 {
1190 	int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
1191 	if (!error)
1192 		validate_coredump_safety();
1193 	return error;
1194 }
1195 
1196 #ifdef CONFIG_COREDUMP
1197 static int proc_dostring_coredump(struct ctl_table *table, int write,
1198 		  void *buffer, size_t *lenp, loff_t *ppos)
1199 {
1200 	int error = proc_dostring(table, write, buffer, lenp, ppos);
1201 	if (!error)
1202 		validate_coredump_safety();
1203 	return error;
1204 }
1205 #endif
1206 
1207 #ifdef CONFIG_MAGIC_SYSRQ
1208 static int sysrq_sysctl_handler(struct ctl_table *table, int write,
1209 				void *buffer, size_t *lenp, loff_t *ppos)
1210 {
1211 	int tmp, ret;
1212 
1213 	tmp = sysrq_mask();
1214 
1215 	ret = __do_proc_dointvec(&tmp, table, write, buffer,
1216 			       lenp, ppos, NULL, NULL);
1217 	if (ret || !write)
1218 		return ret;
1219 
1220 	if (write)
1221 		sysrq_toggle_support(tmp);
1222 
1223 	return 0;
1224 }
1225 #endif
1226 
1227 static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table,
1228 		int write, void *buffer, size_t *lenp, loff_t *ppos,
1229 		unsigned long convmul, unsigned long convdiv)
1230 {
1231 	unsigned long *i, *min, *max;
1232 	int vleft, first = 1, err = 0;
1233 	size_t left;
1234 	char *p;
1235 
1236 	if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
1237 		*lenp = 0;
1238 		return 0;
1239 	}
1240 
1241 	i = (unsigned long *) data;
1242 	min = (unsigned long *) table->extra1;
1243 	max = (unsigned long *) table->extra2;
1244 	vleft = table->maxlen / sizeof(unsigned long);
1245 	left = *lenp;
1246 
1247 	if (write) {
1248 		if (proc_first_pos_non_zero_ignore(ppos, table))
1249 			goto out;
1250 
1251 		if (left > PAGE_SIZE - 1)
1252 			left = PAGE_SIZE - 1;
1253 		p = buffer;
1254 	}
1255 
1256 	for (; left && vleft--; i++, first = 0) {
1257 		unsigned long val;
1258 
1259 		if (write) {
1260 			bool neg;
1261 
1262 			left -= proc_skip_spaces(&p);
1263 			if (!left)
1264 				break;
1265 
1266 			err = proc_get_long(&p, &left, &val, &neg,
1267 					     proc_wspace_sep,
1268 					     sizeof(proc_wspace_sep), NULL);
1269 			if (err)
1270 				break;
1271 			if (neg)
1272 				continue;
1273 			val = convmul * val / convdiv;
1274 			if ((min && val < *min) || (max && val > *max)) {
1275 				err = -EINVAL;
1276 				break;
1277 			}
1278 			*i = val;
1279 		} else {
1280 			val = convdiv * (*i) / convmul;
1281 			if (!first)
1282 				proc_put_char(&buffer, &left, '\t');
1283 			proc_put_long(&buffer, &left, val, false);
1284 		}
1285 	}
1286 
1287 	if (!write && !first && left && !err)
1288 		proc_put_char(&buffer, &left, '\n');
1289 	if (write && !err)
1290 		left -= proc_skip_spaces(&p);
1291 	if (write && first)
1292 		return err ? : -EINVAL;
1293 	*lenp -= left;
1294 out:
1295 	*ppos += *lenp;
1296 	return err;
1297 }
1298 
1299 static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
1300 		void *buffer, size_t *lenp, loff_t *ppos, unsigned long convmul,
1301 		unsigned long convdiv)
1302 {
1303 	return __do_proc_doulongvec_minmax(table->data, table, write,
1304 			buffer, lenp, ppos, convmul, convdiv);
1305 }
1306 
1307 /**
1308  * proc_doulongvec_minmax - read a vector of long integers with min/max values
1309  * @table: the sysctl table
1310  * @write: %TRUE if this is a write to the sysctl file
1311  * @buffer: the user buffer
1312  * @lenp: the size of the user buffer
1313  * @ppos: file position
1314  *
1315  * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1316  * values from/to the user buffer, treated as an ASCII string.
1317  *
1318  * This routine will ensure the values are within the range specified by
1319  * table->extra1 (min) and table->extra2 (max).
1320  *
1321  * Returns 0 on success.
1322  */
1323 int proc_doulongvec_minmax(struct ctl_table *table, int write,
1324 			   void *buffer, size_t *lenp, loff_t *ppos)
1325 {
1326     return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
1327 }
1328 
1329 /**
1330  * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
1331  * @table: the sysctl table
1332  * @write: %TRUE if this is a write to the sysctl file
1333  * @buffer: the user buffer
1334  * @lenp: the size of the user buffer
1335  * @ppos: file position
1336  *
1337  * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1338  * values from/to the user buffer, treated as an ASCII string. The values
1339  * are treated as milliseconds, and converted to jiffies when they are stored.
1340  *
1341  * This routine will ensure the values are within the range specified by
1342  * table->extra1 (min) and table->extra2 (max).
1343  *
1344  * Returns 0 on success.
1345  */
1346 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1347 				      void *buffer, size_t *lenp, loff_t *ppos)
1348 {
1349     return do_proc_doulongvec_minmax(table, write, buffer,
1350 				     lenp, ppos, HZ, 1000l);
1351 }
1352 
1353 
1354 static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
1355 					 int *valp,
1356 					 int write, void *data)
1357 {
1358 	if (write) {
1359 		if (*lvalp > INT_MAX / HZ)
1360 			return 1;
1361 		*valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
1362 	} else {
1363 		int val = *valp;
1364 		unsigned long lval;
1365 		if (val < 0) {
1366 			*negp = true;
1367 			lval = -(unsigned long)val;
1368 		} else {
1369 			*negp = false;
1370 			lval = (unsigned long)val;
1371 		}
1372 		*lvalp = lval / HZ;
1373 	}
1374 	return 0;
1375 }
1376 
1377 static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
1378 						int *valp,
1379 						int write, void *data)
1380 {
1381 	if (write) {
1382 		if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
1383 			return 1;
1384 		*valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
1385 	} else {
1386 		int val = *valp;
1387 		unsigned long lval;
1388 		if (val < 0) {
1389 			*negp = true;
1390 			lval = -(unsigned long)val;
1391 		} else {
1392 			*negp = false;
1393 			lval = (unsigned long)val;
1394 		}
1395 		*lvalp = jiffies_to_clock_t(lval);
1396 	}
1397 	return 0;
1398 }
1399 
1400 static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
1401 					    int *valp,
1402 					    int write, void *data)
1403 {
1404 	if (write) {
1405 		unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
1406 
1407 		if (jif > INT_MAX)
1408 			return 1;
1409 		*valp = (int)jif;
1410 	} else {
1411 		int val = *valp;
1412 		unsigned long lval;
1413 		if (val < 0) {
1414 			*negp = true;
1415 			lval = -(unsigned long)val;
1416 		} else {
1417 			*negp = false;
1418 			lval = (unsigned long)val;
1419 		}
1420 		*lvalp = jiffies_to_msecs(lval);
1421 	}
1422 	return 0;
1423 }
1424 
1425 /**
1426  * proc_dointvec_jiffies - read a vector of integers as seconds
1427  * @table: the sysctl table
1428  * @write: %TRUE if this is a write to the sysctl file
1429  * @buffer: the user buffer
1430  * @lenp: the size of the user buffer
1431  * @ppos: file position
1432  *
1433  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1434  * values from/to the user buffer, treated as an ASCII string.
1435  * The values read are assumed to be in seconds, and are converted into
1436  * jiffies.
1437  *
1438  * Returns 0 on success.
1439  */
1440 int proc_dointvec_jiffies(struct ctl_table *table, int write,
1441 			  void *buffer, size_t *lenp, loff_t *ppos)
1442 {
1443     return do_proc_dointvec(table,write,buffer,lenp,ppos,
1444 		    	    do_proc_dointvec_jiffies_conv,NULL);
1445 }
1446 
1447 /**
1448  * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
1449  * @table: the sysctl table
1450  * @write: %TRUE if this is a write to the sysctl file
1451  * @buffer: the user buffer
1452  * @lenp: the size of the user buffer
1453  * @ppos: pointer to the file position
1454  *
1455  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1456  * values from/to the user buffer, treated as an ASCII string.
1457  * The values read are assumed to be in 1/USER_HZ seconds, and
1458  * are converted into jiffies.
1459  *
1460  * Returns 0 on success.
1461  */
1462 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1463 				 void *buffer, size_t *lenp, loff_t *ppos)
1464 {
1465     return do_proc_dointvec(table,write,buffer,lenp,ppos,
1466 		    	    do_proc_dointvec_userhz_jiffies_conv,NULL);
1467 }
1468 
1469 /**
1470  * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
1471  * @table: the sysctl table
1472  * @write: %TRUE if this is a write to the sysctl file
1473  * @buffer: the user buffer
1474  * @lenp: the size of the user buffer
1475  * @ppos: file position
1476  * @ppos: the current position in the file
1477  *
1478  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1479  * values from/to the user buffer, treated as an ASCII string.
1480  * The values read are assumed to be in 1/1000 seconds, and
1481  * are converted into jiffies.
1482  *
1483  * Returns 0 on success.
1484  */
1485 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write, void *buffer,
1486 		size_t *lenp, loff_t *ppos)
1487 {
1488 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
1489 				do_proc_dointvec_ms_jiffies_conv, NULL);
1490 }
1491 
1492 static int proc_do_cad_pid(struct ctl_table *table, int write, void *buffer,
1493 		size_t *lenp, loff_t *ppos)
1494 {
1495 	struct pid *new_pid;
1496 	pid_t tmp;
1497 	int r;
1498 
1499 	tmp = pid_vnr(cad_pid);
1500 
1501 	r = __do_proc_dointvec(&tmp, table, write, buffer,
1502 			       lenp, ppos, NULL, NULL);
1503 	if (r || !write)
1504 		return r;
1505 
1506 	new_pid = find_get_pid(tmp);
1507 	if (!new_pid)
1508 		return -ESRCH;
1509 
1510 	put_pid(xchg(&cad_pid, new_pid));
1511 	return 0;
1512 }
1513 
1514 /**
1515  * proc_do_large_bitmap - read/write from/to a large bitmap
1516  * @table: the sysctl table
1517  * @write: %TRUE if this is a write to the sysctl file
1518  * @buffer: the user buffer
1519  * @lenp: the size of the user buffer
1520  * @ppos: file position
1521  *
1522  * The bitmap is stored at table->data and the bitmap length (in bits)
1523  * in table->maxlen.
1524  *
1525  * We use a range comma separated format (e.g. 1,3-4,10-10) so that
1526  * large bitmaps may be represented in a compact manner. Writing into
1527  * the file will clear the bitmap then update it with the given input.
1528  *
1529  * Returns 0 on success.
1530  */
1531 int proc_do_large_bitmap(struct ctl_table *table, int write,
1532 			 void *buffer, size_t *lenp, loff_t *ppos)
1533 {
1534 	int err = 0;
1535 	size_t left = *lenp;
1536 	unsigned long bitmap_len = table->maxlen;
1537 	unsigned long *bitmap = *(unsigned long **) table->data;
1538 	unsigned long *tmp_bitmap = NULL;
1539 	char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
1540 
1541 	if (!bitmap || !bitmap_len || !left || (*ppos && !write)) {
1542 		*lenp = 0;
1543 		return 0;
1544 	}
1545 
1546 	if (write) {
1547 		char *p = buffer;
1548 		size_t skipped = 0;
1549 
1550 		if (left > PAGE_SIZE - 1) {
1551 			left = PAGE_SIZE - 1;
1552 			/* How much of the buffer we'll skip this pass */
1553 			skipped = *lenp - left;
1554 		}
1555 
1556 		tmp_bitmap = bitmap_zalloc(bitmap_len, GFP_KERNEL);
1557 		if (!tmp_bitmap)
1558 			return -ENOMEM;
1559 		proc_skip_char(&p, &left, '\n');
1560 		while (!err && left) {
1561 			unsigned long val_a, val_b;
1562 			bool neg;
1563 			size_t saved_left;
1564 
1565 			/* In case we stop parsing mid-number, we can reset */
1566 			saved_left = left;
1567 			err = proc_get_long(&p, &left, &val_a, &neg, tr_a,
1568 					     sizeof(tr_a), &c);
1569 			/*
1570 			 * If we consumed the entirety of a truncated buffer or
1571 			 * only one char is left (may be a "-"), then stop here,
1572 			 * reset, & come back for more.
1573 			 */
1574 			if ((left <= 1) && skipped) {
1575 				left = saved_left;
1576 				break;
1577 			}
1578 
1579 			if (err)
1580 				break;
1581 			if (val_a >= bitmap_len || neg) {
1582 				err = -EINVAL;
1583 				break;
1584 			}
1585 
1586 			val_b = val_a;
1587 			if (left) {
1588 				p++;
1589 				left--;
1590 			}
1591 
1592 			if (c == '-') {
1593 				err = proc_get_long(&p, &left, &val_b,
1594 						     &neg, tr_b, sizeof(tr_b),
1595 						     &c);
1596 				/*
1597 				 * If we consumed all of a truncated buffer or
1598 				 * then stop here, reset, & come back for more.
1599 				 */
1600 				if (!left && skipped) {
1601 					left = saved_left;
1602 					break;
1603 				}
1604 
1605 				if (err)
1606 					break;
1607 				if (val_b >= bitmap_len || neg ||
1608 				    val_a > val_b) {
1609 					err = -EINVAL;
1610 					break;
1611 				}
1612 				if (left) {
1613 					p++;
1614 					left--;
1615 				}
1616 			}
1617 
1618 			bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
1619 			proc_skip_char(&p, &left, '\n');
1620 		}
1621 		left += skipped;
1622 	} else {
1623 		unsigned long bit_a, bit_b = 0;
1624 		bool first = 1;
1625 
1626 		while (left) {
1627 			bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
1628 			if (bit_a >= bitmap_len)
1629 				break;
1630 			bit_b = find_next_zero_bit(bitmap, bitmap_len,
1631 						   bit_a + 1) - 1;
1632 
1633 			if (!first)
1634 				proc_put_char(&buffer, &left, ',');
1635 			proc_put_long(&buffer, &left, bit_a, false);
1636 			if (bit_a != bit_b) {
1637 				proc_put_char(&buffer, &left, '-');
1638 				proc_put_long(&buffer, &left, bit_b, false);
1639 			}
1640 
1641 			first = 0; bit_b++;
1642 		}
1643 		proc_put_char(&buffer, &left, '\n');
1644 	}
1645 
1646 	if (!err) {
1647 		if (write) {
1648 			if (*ppos)
1649 				bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
1650 			else
1651 				bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
1652 		}
1653 		*lenp -= left;
1654 		*ppos += *lenp;
1655 	}
1656 
1657 	bitmap_free(tmp_bitmap);
1658 	return err;
1659 }
1660 
1661 #else /* CONFIG_PROC_SYSCTL */
1662 
1663 int proc_dostring(struct ctl_table *table, int write,
1664 		  void *buffer, size_t *lenp, loff_t *ppos)
1665 {
1666 	return -ENOSYS;
1667 }
1668 
1669 int proc_dobool(struct ctl_table *table, int write,
1670 		void *buffer, size_t *lenp, loff_t *ppos)
1671 {
1672 	return -ENOSYS;
1673 }
1674 
1675 int proc_dointvec(struct ctl_table *table, int write,
1676 		  void *buffer, size_t *lenp, loff_t *ppos)
1677 {
1678 	return -ENOSYS;
1679 }
1680 
1681 int proc_douintvec(struct ctl_table *table, int write,
1682 		  void *buffer, size_t *lenp, loff_t *ppos)
1683 {
1684 	return -ENOSYS;
1685 }
1686 
1687 int proc_dointvec_minmax(struct ctl_table *table, int write,
1688 		    void *buffer, size_t *lenp, loff_t *ppos)
1689 {
1690 	return -ENOSYS;
1691 }
1692 
1693 int proc_douintvec_minmax(struct ctl_table *table, int write,
1694 			  void *buffer, size_t *lenp, loff_t *ppos)
1695 {
1696 	return -ENOSYS;
1697 }
1698 
1699 int proc_dou8vec_minmax(struct ctl_table *table, int write,
1700 			void *buffer, size_t *lenp, loff_t *ppos)
1701 {
1702 	return -ENOSYS;
1703 }
1704 
1705 int proc_dointvec_jiffies(struct ctl_table *table, int write,
1706 		    void *buffer, size_t *lenp, loff_t *ppos)
1707 {
1708 	return -ENOSYS;
1709 }
1710 
1711 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1712 		    void *buffer, size_t *lenp, loff_t *ppos)
1713 {
1714 	return -ENOSYS;
1715 }
1716 
1717 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
1718 			     void *buffer, size_t *lenp, loff_t *ppos)
1719 {
1720 	return -ENOSYS;
1721 }
1722 
1723 int proc_doulongvec_minmax(struct ctl_table *table, int write,
1724 		    void *buffer, size_t *lenp, loff_t *ppos)
1725 {
1726 	return -ENOSYS;
1727 }
1728 
1729 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1730 				      void *buffer, size_t *lenp, loff_t *ppos)
1731 {
1732 	return -ENOSYS;
1733 }
1734 
1735 int proc_do_large_bitmap(struct ctl_table *table, int write,
1736 			 void *buffer, size_t *lenp, loff_t *ppos)
1737 {
1738 	return -ENOSYS;
1739 }
1740 
1741 #endif /* CONFIG_PROC_SYSCTL */
1742 
1743 #if defined(CONFIG_SYSCTL)
1744 int proc_do_static_key(struct ctl_table *table, int write,
1745 		       void *buffer, size_t *lenp, loff_t *ppos)
1746 {
1747 	struct static_key *key = (struct static_key *)table->data;
1748 	static DEFINE_MUTEX(static_key_mutex);
1749 	int val, ret;
1750 	struct ctl_table tmp = {
1751 		.data   = &val,
1752 		.maxlen = sizeof(val),
1753 		.mode   = table->mode,
1754 		.extra1 = SYSCTL_ZERO,
1755 		.extra2 = SYSCTL_ONE,
1756 	};
1757 
1758 	if (write && !capable(CAP_SYS_ADMIN))
1759 		return -EPERM;
1760 
1761 	mutex_lock(&static_key_mutex);
1762 	val = static_key_enabled(key);
1763 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
1764 	if (write && !ret) {
1765 		if (val)
1766 			static_key_enable(key);
1767 		else
1768 			static_key_disable(key);
1769 	}
1770 	mutex_unlock(&static_key_mutex);
1771 	return ret;
1772 }
1773 
1774 static struct ctl_table kern_table[] = {
1775 	{
1776 		.procname	= "sched_child_runs_first",
1777 		.data		= &sysctl_sched_child_runs_first,
1778 		.maxlen		= sizeof(unsigned int),
1779 		.mode		= 0644,
1780 		.proc_handler	= proc_dointvec,
1781 	},
1782 #ifdef CONFIG_SCHEDSTATS
1783 	{
1784 		.procname	= "sched_schedstats",
1785 		.data		= NULL,
1786 		.maxlen		= sizeof(unsigned int),
1787 		.mode		= 0644,
1788 		.proc_handler	= sysctl_schedstats,
1789 		.extra1		= SYSCTL_ZERO,
1790 		.extra2		= SYSCTL_ONE,
1791 	},
1792 #endif /* CONFIG_SCHEDSTATS */
1793 #ifdef CONFIG_TASK_DELAY_ACCT
1794 	{
1795 		.procname	= "task_delayacct",
1796 		.data		= NULL,
1797 		.maxlen		= sizeof(unsigned int),
1798 		.mode		= 0644,
1799 		.proc_handler	= sysctl_delayacct,
1800 		.extra1		= SYSCTL_ZERO,
1801 		.extra2		= SYSCTL_ONE,
1802 	},
1803 #endif /* CONFIG_TASK_DELAY_ACCT */
1804 #ifdef CONFIG_NUMA_BALANCING
1805 	{
1806 		.procname	= "numa_balancing",
1807 		.data		= NULL, /* filled in by handler */
1808 		.maxlen		= sizeof(unsigned int),
1809 		.mode		= 0644,
1810 		.proc_handler	= sysctl_numa_balancing,
1811 		.extra1		= SYSCTL_ZERO,
1812 		.extra2		= SYSCTL_ONE,
1813 	},
1814 #endif /* CONFIG_NUMA_BALANCING */
1815 	{
1816 		.procname	= "sched_rt_period_us",
1817 		.data		= &sysctl_sched_rt_period,
1818 		.maxlen		= sizeof(unsigned int),
1819 		.mode		= 0644,
1820 		.proc_handler	= sched_rt_handler,
1821 	},
1822 	{
1823 		.procname	= "sched_rt_runtime_us",
1824 		.data		= &sysctl_sched_rt_runtime,
1825 		.maxlen		= sizeof(int),
1826 		.mode		= 0644,
1827 		.proc_handler	= sched_rt_handler,
1828 	},
1829 	{
1830 		.procname	= "sched_deadline_period_max_us",
1831 		.data		= &sysctl_sched_dl_period_max,
1832 		.maxlen		= sizeof(unsigned int),
1833 		.mode		= 0644,
1834 		.proc_handler	= proc_dointvec,
1835 	},
1836 	{
1837 		.procname	= "sched_deadline_period_min_us",
1838 		.data		= &sysctl_sched_dl_period_min,
1839 		.maxlen		= sizeof(unsigned int),
1840 		.mode		= 0644,
1841 		.proc_handler	= proc_dointvec,
1842 	},
1843 	{
1844 		.procname	= "sched_rr_timeslice_ms",
1845 		.data		= &sysctl_sched_rr_timeslice,
1846 		.maxlen		= sizeof(int),
1847 		.mode		= 0644,
1848 		.proc_handler	= sched_rr_handler,
1849 	},
1850 #ifdef CONFIG_UCLAMP_TASK
1851 	{
1852 		.procname	= "sched_util_clamp_min",
1853 		.data		= &sysctl_sched_uclamp_util_min,
1854 		.maxlen		= sizeof(unsigned int),
1855 		.mode		= 0644,
1856 		.proc_handler	= sysctl_sched_uclamp_handler,
1857 	},
1858 	{
1859 		.procname	= "sched_util_clamp_max",
1860 		.data		= &sysctl_sched_uclamp_util_max,
1861 		.maxlen		= sizeof(unsigned int),
1862 		.mode		= 0644,
1863 		.proc_handler	= sysctl_sched_uclamp_handler,
1864 	},
1865 	{
1866 		.procname	= "sched_util_clamp_min_rt_default",
1867 		.data		= &sysctl_sched_uclamp_util_min_rt_default,
1868 		.maxlen		= sizeof(unsigned int),
1869 		.mode		= 0644,
1870 		.proc_handler	= sysctl_sched_uclamp_handler,
1871 	},
1872 #endif
1873 #ifdef CONFIG_SCHED_AUTOGROUP
1874 	{
1875 		.procname	= "sched_autogroup_enabled",
1876 		.data		= &sysctl_sched_autogroup_enabled,
1877 		.maxlen		= sizeof(unsigned int),
1878 		.mode		= 0644,
1879 		.proc_handler	= proc_dointvec_minmax,
1880 		.extra1		= SYSCTL_ZERO,
1881 		.extra2		= SYSCTL_ONE,
1882 	},
1883 #endif
1884 #ifdef CONFIG_CFS_BANDWIDTH
1885 	{
1886 		.procname	= "sched_cfs_bandwidth_slice_us",
1887 		.data		= &sysctl_sched_cfs_bandwidth_slice,
1888 		.maxlen		= sizeof(unsigned int),
1889 		.mode		= 0644,
1890 		.proc_handler	= proc_dointvec_minmax,
1891 		.extra1		= SYSCTL_ONE,
1892 	},
1893 #endif
1894 #if defined(CONFIG_ENERGY_MODEL) && defined(CONFIG_CPU_FREQ_GOV_SCHEDUTIL)
1895 	{
1896 		.procname	= "sched_energy_aware",
1897 		.data		= &sysctl_sched_energy_aware,
1898 		.maxlen		= sizeof(unsigned int),
1899 		.mode		= 0644,
1900 		.proc_handler	= sched_energy_aware_handler,
1901 		.extra1		= SYSCTL_ZERO,
1902 		.extra2		= SYSCTL_ONE,
1903 	},
1904 #endif
1905 #ifdef CONFIG_PROVE_LOCKING
1906 	{
1907 		.procname	= "prove_locking",
1908 		.data		= &prove_locking,
1909 		.maxlen		= sizeof(int),
1910 		.mode		= 0644,
1911 		.proc_handler	= proc_dointvec,
1912 	},
1913 #endif
1914 #ifdef CONFIG_LOCK_STAT
1915 	{
1916 		.procname	= "lock_stat",
1917 		.data		= &lock_stat,
1918 		.maxlen		= sizeof(int),
1919 		.mode		= 0644,
1920 		.proc_handler	= proc_dointvec,
1921 	},
1922 #endif
1923 	{
1924 		.procname	= "panic",
1925 		.data		= &panic_timeout,
1926 		.maxlen		= sizeof(int),
1927 		.mode		= 0644,
1928 		.proc_handler	= proc_dointvec,
1929 	},
1930 #ifdef CONFIG_COREDUMP
1931 	{
1932 		.procname	= "core_uses_pid",
1933 		.data		= &core_uses_pid,
1934 		.maxlen		= sizeof(int),
1935 		.mode		= 0644,
1936 		.proc_handler	= proc_dointvec,
1937 	},
1938 	{
1939 		.procname	= "core_pattern",
1940 		.data		= core_pattern,
1941 		.maxlen		= CORENAME_MAX_SIZE,
1942 		.mode		= 0644,
1943 		.proc_handler	= proc_dostring_coredump,
1944 	},
1945 	{
1946 		.procname	= "core_pipe_limit",
1947 		.data		= &core_pipe_limit,
1948 		.maxlen		= sizeof(unsigned int),
1949 		.mode		= 0644,
1950 		.proc_handler	= proc_dointvec,
1951 	},
1952 #endif
1953 #ifdef CONFIG_PROC_SYSCTL
1954 	{
1955 		.procname	= "tainted",
1956 		.maxlen 	= sizeof(long),
1957 		.mode		= 0644,
1958 		.proc_handler	= proc_taint,
1959 	},
1960 	{
1961 		.procname	= "sysctl_writes_strict",
1962 		.data		= &sysctl_writes_strict,
1963 		.maxlen		= sizeof(int),
1964 		.mode		= 0644,
1965 		.proc_handler	= proc_dointvec_minmax,
1966 		.extra1		= &neg_one,
1967 		.extra2		= SYSCTL_ONE,
1968 	},
1969 #endif
1970 #ifdef CONFIG_LATENCYTOP
1971 	{
1972 		.procname	= "latencytop",
1973 		.data		= &latencytop_enabled,
1974 		.maxlen		= sizeof(int),
1975 		.mode		= 0644,
1976 		.proc_handler	= sysctl_latencytop,
1977 	},
1978 #endif
1979 #ifdef CONFIG_BLK_DEV_INITRD
1980 	{
1981 		.procname	= "real-root-dev",
1982 		.data		= &real_root_dev,
1983 		.maxlen		= sizeof(int),
1984 		.mode		= 0644,
1985 		.proc_handler	= proc_dointvec,
1986 	},
1987 #endif
1988 	{
1989 		.procname	= "print-fatal-signals",
1990 		.data		= &print_fatal_signals,
1991 		.maxlen		= sizeof(int),
1992 		.mode		= 0644,
1993 		.proc_handler	= proc_dointvec,
1994 	},
1995 #ifdef CONFIG_SPARC
1996 	{
1997 		.procname	= "reboot-cmd",
1998 		.data		= reboot_command,
1999 		.maxlen		= 256,
2000 		.mode		= 0644,
2001 		.proc_handler	= proc_dostring,
2002 	},
2003 	{
2004 		.procname	= "stop-a",
2005 		.data		= &stop_a_enabled,
2006 		.maxlen		= sizeof (int),
2007 		.mode		= 0644,
2008 		.proc_handler	= proc_dointvec,
2009 	},
2010 	{
2011 		.procname	= "scons-poweroff",
2012 		.data		= &scons_pwroff,
2013 		.maxlen		= sizeof (int),
2014 		.mode		= 0644,
2015 		.proc_handler	= proc_dointvec,
2016 	},
2017 #endif
2018 #ifdef CONFIG_SPARC64
2019 	{
2020 		.procname	= "tsb-ratio",
2021 		.data		= &sysctl_tsb_ratio,
2022 		.maxlen		= sizeof (int),
2023 		.mode		= 0644,
2024 		.proc_handler	= proc_dointvec,
2025 	},
2026 #endif
2027 #ifdef CONFIG_PARISC
2028 	{
2029 		.procname	= "soft-power",
2030 		.data		= &pwrsw_enabled,
2031 		.maxlen		= sizeof (int),
2032 		.mode		= 0644,
2033 		.proc_handler	= proc_dointvec,
2034 	},
2035 #endif
2036 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
2037 	{
2038 		.procname	= "unaligned-trap",
2039 		.data		= &unaligned_enabled,
2040 		.maxlen		= sizeof (int),
2041 		.mode		= 0644,
2042 		.proc_handler	= proc_dointvec,
2043 	},
2044 #endif
2045 	{
2046 		.procname	= "ctrl-alt-del",
2047 		.data		= &C_A_D,
2048 		.maxlen		= sizeof(int),
2049 		.mode		= 0644,
2050 		.proc_handler	= proc_dointvec,
2051 	},
2052 #ifdef CONFIG_FUNCTION_TRACER
2053 	{
2054 		.procname	= "ftrace_enabled",
2055 		.data		= &ftrace_enabled,
2056 		.maxlen		= sizeof(int),
2057 		.mode		= 0644,
2058 		.proc_handler	= ftrace_enable_sysctl,
2059 	},
2060 #endif
2061 #ifdef CONFIG_STACK_TRACER
2062 	{
2063 		.procname	= "stack_tracer_enabled",
2064 		.data		= &stack_tracer_enabled,
2065 		.maxlen		= sizeof(int),
2066 		.mode		= 0644,
2067 		.proc_handler	= stack_trace_sysctl,
2068 	},
2069 #endif
2070 #ifdef CONFIG_TRACING
2071 	{
2072 		.procname	= "ftrace_dump_on_oops",
2073 		.data		= &ftrace_dump_on_oops,
2074 		.maxlen		= sizeof(int),
2075 		.mode		= 0644,
2076 		.proc_handler	= proc_dointvec,
2077 	},
2078 	{
2079 		.procname	= "traceoff_on_warning",
2080 		.data		= &__disable_trace_on_warning,
2081 		.maxlen		= sizeof(__disable_trace_on_warning),
2082 		.mode		= 0644,
2083 		.proc_handler	= proc_dointvec,
2084 	},
2085 	{
2086 		.procname	= "tracepoint_printk",
2087 		.data		= &tracepoint_printk,
2088 		.maxlen		= sizeof(tracepoint_printk),
2089 		.mode		= 0644,
2090 		.proc_handler	= tracepoint_printk_sysctl,
2091 	},
2092 #endif
2093 #ifdef CONFIG_KEXEC_CORE
2094 	{
2095 		.procname	= "kexec_load_disabled",
2096 		.data		= &kexec_load_disabled,
2097 		.maxlen		= sizeof(int),
2098 		.mode		= 0644,
2099 		/* only handle a transition from default "0" to "1" */
2100 		.proc_handler	= proc_dointvec_minmax,
2101 		.extra1		= SYSCTL_ONE,
2102 		.extra2		= SYSCTL_ONE,
2103 	},
2104 #endif
2105 #ifdef CONFIG_MODULES
2106 	{
2107 		.procname	= "modprobe",
2108 		.data		= &modprobe_path,
2109 		.maxlen		= KMOD_PATH_LEN,
2110 		.mode		= 0644,
2111 		.proc_handler	= proc_dostring,
2112 	},
2113 	{
2114 		.procname	= "modules_disabled",
2115 		.data		= &modules_disabled,
2116 		.maxlen		= sizeof(int),
2117 		.mode		= 0644,
2118 		/* only handle a transition from default "0" to "1" */
2119 		.proc_handler	= proc_dointvec_minmax,
2120 		.extra1		= SYSCTL_ONE,
2121 		.extra2		= SYSCTL_ONE,
2122 	},
2123 #endif
2124 #ifdef CONFIG_UEVENT_HELPER
2125 	{
2126 		.procname	= "hotplug",
2127 		.data		= &uevent_helper,
2128 		.maxlen		= UEVENT_HELPER_PATH_LEN,
2129 		.mode		= 0644,
2130 		.proc_handler	= proc_dostring,
2131 	},
2132 #endif
2133 #ifdef CONFIG_CHR_DEV_SG
2134 	{
2135 		.procname	= "sg-big-buff",
2136 		.data		= &sg_big_buff,
2137 		.maxlen		= sizeof (int),
2138 		.mode		= 0444,
2139 		.proc_handler	= proc_dointvec,
2140 	},
2141 #endif
2142 #ifdef CONFIG_BSD_PROCESS_ACCT
2143 	{
2144 		.procname	= "acct",
2145 		.data		= &acct_parm,
2146 		.maxlen		= 3*sizeof(int),
2147 		.mode		= 0644,
2148 		.proc_handler	= proc_dointvec,
2149 	},
2150 #endif
2151 #ifdef CONFIG_MAGIC_SYSRQ
2152 	{
2153 		.procname	= "sysrq",
2154 		.data		= NULL,
2155 		.maxlen		= sizeof (int),
2156 		.mode		= 0644,
2157 		.proc_handler	= sysrq_sysctl_handler,
2158 	},
2159 #endif
2160 #ifdef CONFIG_PROC_SYSCTL
2161 	{
2162 		.procname	= "cad_pid",
2163 		.data		= NULL,
2164 		.maxlen		= sizeof (int),
2165 		.mode		= 0600,
2166 		.proc_handler	= proc_do_cad_pid,
2167 	},
2168 #endif
2169 	{
2170 		.procname	= "threads-max",
2171 		.data		= NULL,
2172 		.maxlen		= sizeof(int),
2173 		.mode		= 0644,
2174 		.proc_handler	= sysctl_max_threads,
2175 	},
2176 	{
2177 		.procname	= "random",
2178 		.mode		= 0555,
2179 		.child		= random_table,
2180 	},
2181 	{
2182 		.procname	= "usermodehelper",
2183 		.mode		= 0555,
2184 		.child		= usermodehelper_table,
2185 	},
2186 #ifdef CONFIG_FW_LOADER_USER_HELPER
2187 	{
2188 		.procname	= "firmware_config",
2189 		.mode		= 0555,
2190 		.child		= firmware_config_table,
2191 	},
2192 #endif
2193 	{
2194 		.procname	= "overflowuid",
2195 		.data		= &overflowuid,
2196 		.maxlen		= sizeof(int),
2197 		.mode		= 0644,
2198 		.proc_handler	= proc_dointvec_minmax,
2199 		.extra1		= &minolduid,
2200 		.extra2		= &maxolduid,
2201 	},
2202 	{
2203 		.procname	= "overflowgid",
2204 		.data		= &overflowgid,
2205 		.maxlen		= sizeof(int),
2206 		.mode		= 0644,
2207 		.proc_handler	= proc_dointvec_minmax,
2208 		.extra1		= &minolduid,
2209 		.extra2		= &maxolduid,
2210 	},
2211 #ifdef CONFIG_S390
2212 	{
2213 		.procname	= "userprocess_debug",
2214 		.data		= &show_unhandled_signals,
2215 		.maxlen		= sizeof(int),
2216 		.mode		= 0644,
2217 		.proc_handler	= proc_dointvec,
2218 	},
2219 #endif
2220 #ifdef CONFIG_SMP
2221 	{
2222 		.procname	= "oops_all_cpu_backtrace",
2223 		.data		= &sysctl_oops_all_cpu_backtrace,
2224 		.maxlen		= sizeof(int),
2225 		.mode		= 0644,
2226 		.proc_handler	= proc_dointvec_minmax,
2227 		.extra1		= SYSCTL_ZERO,
2228 		.extra2		= SYSCTL_ONE,
2229 	},
2230 #endif /* CONFIG_SMP */
2231 	{
2232 		.procname	= "pid_max",
2233 		.data		= &pid_max,
2234 		.maxlen		= sizeof (int),
2235 		.mode		= 0644,
2236 		.proc_handler	= proc_dointvec_minmax,
2237 		.extra1		= &pid_max_min,
2238 		.extra2		= &pid_max_max,
2239 	},
2240 	{
2241 		.procname	= "panic_on_oops",
2242 		.data		= &panic_on_oops,
2243 		.maxlen		= sizeof(int),
2244 		.mode		= 0644,
2245 		.proc_handler	= proc_dointvec,
2246 	},
2247 	{
2248 		.procname	= "panic_print",
2249 		.data		= &panic_print,
2250 		.maxlen		= sizeof(unsigned long),
2251 		.mode		= 0644,
2252 		.proc_handler	= proc_doulongvec_minmax,
2253 	},
2254 #if defined CONFIG_PRINTK
2255 	{
2256 		.procname	= "printk",
2257 		.data		= &console_loglevel,
2258 		.maxlen		= 4*sizeof(int),
2259 		.mode		= 0644,
2260 		.proc_handler	= proc_dointvec,
2261 	},
2262 	{
2263 		.procname	= "printk_ratelimit",
2264 		.data		= &printk_ratelimit_state.interval,
2265 		.maxlen		= sizeof(int),
2266 		.mode		= 0644,
2267 		.proc_handler	= proc_dointvec_jiffies,
2268 	},
2269 	{
2270 		.procname	= "printk_ratelimit_burst",
2271 		.data		= &printk_ratelimit_state.burst,
2272 		.maxlen		= sizeof(int),
2273 		.mode		= 0644,
2274 		.proc_handler	= proc_dointvec,
2275 	},
2276 	{
2277 		.procname	= "printk_delay",
2278 		.data		= &printk_delay_msec,
2279 		.maxlen		= sizeof(int),
2280 		.mode		= 0644,
2281 		.proc_handler	= proc_dointvec_minmax,
2282 		.extra1		= SYSCTL_ZERO,
2283 		.extra2		= &ten_thousand,
2284 	},
2285 	{
2286 		.procname	= "printk_devkmsg",
2287 		.data		= devkmsg_log_str,
2288 		.maxlen		= DEVKMSG_STR_MAX_SIZE,
2289 		.mode		= 0644,
2290 		.proc_handler	= devkmsg_sysctl_set_loglvl,
2291 	},
2292 	{
2293 		.procname	= "dmesg_restrict",
2294 		.data		= &dmesg_restrict,
2295 		.maxlen		= sizeof(int),
2296 		.mode		= 0644,
2297 		.proc_handler	= proc_dointvec_minmax_sysadmin,
2298 		.extra1		= SYSCTL_ZERO,
2299 		.extra2		= SYSCTL_ONE,
2300 	},
2301 	{
2302 		.procname	= "kptr_restrict",
2303 		.data		= &kptr_restrict,
2304 		.maxlen		= sizeof(int),
2305 		.mode		= 0644,
2306 		.proc_handler	= proc_dointvec_minmax_sysadmin,
2307 		.extra1		= SYSCTL_ZERO,
2308 		.extra2		= &two,
2309 	},
2310 #endif
2311 	{
2312 		.procname	= "ngroups_max",
2313 		.data		= &ngroups_max,
2314 		.maxlen		= sizeof (int),
2315 		.mode		= 0444,
2316 		.proc_handler	= proc_dointvec,
2317 	},
2318 	{
2319 		.procname	= "cap_last_cap",
2320 		.data		= (void *)&cap_last_cap,
2321 		.maxlen		= sizeof(int),
2322 		.mode		= 0444,
2323 		.proc_handler	= proc_dointvec,
2324 	},
2325 #if defined(CONFIG_LOCKUP_DETECTOR)
2326 	{
2327 		.procname       = "watchdog",
2328 		.data		= &watchdog_user_enabled,
2329 		.maxlen		= sizeof(int),
2330 		.mode		= 0644,
2331 		.proc_handler   = proc_watchdog,
2332 		.extra1		= SYSCTL_ZERO,
2333 		.extra2		= SYSCTL_ONE,
2334 	},
2335 	{
2336 		.procname	= "watchdog_thresh",
2337 		.data		= &watchdog_thresh,
2338 		.maxlen		= sizeof(int),
2339 		.mode		= 0644,
2340 		.proc_handler	= proc_watchdog_thresh,
2341 		.extra1		= SYSCTL_ZERO,
2342 		.extra2		= &sixty,
2343 	},
2344 	{
2345 		.procname       = "nmi_watchdog",
2346 		.data		= &nmi_watchdog_user_enabled,
2347 		.maxlen		= sizeof(int),
2348 		.mode		= NMI_WATCHDOG_SYSCTL_PERM,
2349 		.proc_handler   = proc_nmi_watchdog,
2350 		.extra1		= SYSCTL_ZERO,
2351 		.extra2		= SYSCTL_ONE,
2352 	},
2353 	{
2354 		.procname	= "watchdog_cpumask",
2355 		.data		= &watchdog_cpumask_bits,
2356 		.maxlen		= NR_CPUS,
2357 		.mode		= 0644,
2358 		.proc_handler	= proc_watchdog_cpumask,
2359 	},
2360 #ifdef CONFIG_SOFTLOCKUP_DETECTOR
2361 	{
2362 		.procname       = "soft_watchdog",
2363 		.data		= &soft_watchdog_user_enabled,
2364 		.maxlen		= sizeof(int),
2365 		.mode		= 0644,
2366 		.proc_handler   = proc_soft_watchdog,
2367 		.extra1		= SYSCTL_ZERO,
2368 		.extra2		= SYSCTL_ONE,
2369 	},
2370 	{
2371 		.procname	= "softlockup_panic",
2372 		.data		= &softlockup_panic,
2373 		.maxlen		= sizeof(int),
2374 		.mode		= 0644,
2375 		.proc_handler	= proc_dointvec_minmax,
2376 		.extra1		= SYSCTL_ZERO,
2377 		.extra2		= SYSCTL_ONE,
2378 	},
2379 #ifdef CONFIG_SMP
2380 	{
2381 		.procname	= "softlockup_all_cpu_backtrace",
2382 		.data		= &sysctl_softlockup_all_cpu_backtrace,
2383 		.maxlen		= sizeof(int),
2384 		.mode		= 0644,
2385 		.proc_handler	= proc_dointvec_minmax,
2386 		.extra1		= SYSCTL_ZERO,
2387 		.extra2		= SYSCTL_ONE,
2388 	},
2389 #endif /* CONFIG_SMP */
2390 #endif
2391 #ifdef CONFIG_HARDLOCKUP_DETECTOR
2392 	{
2393 		.procname	= "hardlockup_panic",
2394 		.data		= &hardlockup_panic,
2395 		.maxlen		= sizeof(int),
2396 		.mode		= 0644,
2397 		.proc_handler	= proc_dointvec_minmax,
2398 		.extra1		= SYSCTL_ZERO,
2399 		.extra2		= SYSCTL_ONE,
2400 	},
2401 #ifdef CONFIG_SMP
2402 	{
2403 		.procname	= "hardlockup_all_cpu_backtrace",
2404 		.data		= &sysctl_hardlockup_all_cpu_backtrace,
2405 		.maxlen		= sizeof(int),
2406 		.mode		= 0644,
2407 		.proc_handler	= proc_dointvec_minmax,
2408 		.extra1		= SYSCTL_ZERO,
2409 		.extra2		= SYSCTL_ONE,
2410 	},
2411 #endif /* CONFIG_SMP */
2412 #endif
2413 #endif
2414 
2415 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
2416 	{
2417 		.procname       = "unknown_nmi_panic",
2418 		.data           = &unknown_nmi_panic,
2419 		.maxlen         = sizeof (int),
2420 		.mode           = 0644,
2421 		.proc_handler   = proc_dointvec,
2422 	},
2423 #endif
2424 
2425 #if (defined(CONFIG_X86_32) || defined(CONFIG_PARISC)) && \
2426 	defined(CONFIG_DEBUG_STACKOVERFLOW)
2427 	{
2428 		.procname	= "panic_on_stackoverflow",
2429 		.data		= &sysctl_panic_on_stackoverflow,
2430 		.maxlen		= sizeof(int),
2431 		.mode		= 0644,
2432 		.proc_handler	= proc_dointvec,
2433 	},
2434 #endif
2435 #if defined(CONFIG_X86)
2436 	{
2437 		.procname	= "panic_on_unrecovered_nmi",
2438 		.data		= &panic_on_unrecovered_nmi,
2439 		.maxlen		= sizeof(int),
2440 		.mode		= 0644,
2441 		.proc_handler	= proc_dointvec,
2442 	},
2443 	{
2444 		.procname	= "panic_on_io_nmi",
2445 		.data		= &panic_on_io_nmi,
2446 		.maxlen		= sizeof(int),
2447 		.mode		= 0644,
2448 		.proc_handler	= proc_dointvec,
2449 	},
2450 	{
2451 		.procname	= "bootloader_type",
2452 		.data		= &bootloader_type,
2453 		.maxlen		= sizeof (int),
2454 		.mode		= 0444,
2455 		.proc_handler	= proc_dointvec,
2456 	},
2457 	{
2458 		.procname	= "bootloader_version",
2459 		.data		= &bootloader_version,
2460 		.maxlen		= sizeof (int),
2461 		.mode		= 0444,
2462 		.proc_handler	= proc_dointvec,
2463 	},
2464 	{
2465 		.procname	= "io_delay_type",
2466 		.data		= &io_delay_type,
2467 		.maxlen		= sizeof(int),
2468 		.mode		= 0644,
2469 		.proc_handler	= proc_dointvec,
2470 	},
2471 #endif
2472 #if defined(CONFIG_MMU)
2473 	{
2474 		.procname	= "randomize_va_space",
2475 		.data		= &randomize_va_space,
2476 		.maxlen		= sizeof(int),
2477 		.mode		= 0644,
2478 		.proc_handler	= proc_dointvec,
2479 	},
2480 #endif
2481 #if defined(CONFIG_S390) && defined(CONFIG_SMP)
2482 	{
2483 		.procname	= "spin_retry",
2484 		.data		= &spin_retry,
2485 		.maxlen		= sizeof (int),
2486 		.mode		= 0644,
2487 		.proc_handler	= proc_dointvec,
2488 	},
2489 #endif
2490 #if	defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
2491 	{
2492 		.procname	= "acpi_video_flags",
2493 		.data		= &acpi_realmode_flags,
2494 		.maxlen		= sizeof (unsigned long),
2495 		.mode		= 0644,
2496 		.proc_handler	= proc_doulongvec_minmax,
2497 	},
2498 #endif
2499 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
2500 	{
2501 		.procname	= "ignore-unaligned-usertrap",
2502 		.data		= &no_unaligned_warning,
2503 		.maxlen		= sizeof (int),
2504 		.mode		= 0644,
2505 		.proc_handler	= proc_dointvec,
2506 	},
2507 #endif
2508 #ifdef CONFIG_IA64
2509 	{
2510 		.procname	= "unaligned-dump-stack",
2511 		.data		= &unaligned_dump_stack,
2512 		.maxlen		= sizeof (int),
2513 		.mode		= 0644,
2514 		.proc_handler	= proc_dointvec,
2515 	},
2516 #endif
2517 #ifdef CONFIG_DETECT_HUNG_TASK
2518 #ifdef CONFIG_SMP
2519 	{
2520 		.procname	= "hung_task_all_cpu_backtrace",
2521 		.data		= &sysctl_hung_task_all_cpu_backtrace,
2522 		.maxlen		= sizeof(int),
2523 		.mode		= 0644,
2524 		.proc_handler	= proc_dointvec_minmax,
2525 		.extra1		= SYSCTL_ZERO,
2526 		.extra2		= SYSCTL_ONE,
2527 	},
2528 #endif /* CONFIG_SMP */
2529 	{
2530 		.procname	= "hung_task_panic",
2531 		.data		= &sysctl_hung_task_panic,
2532 		.maxlen		= sizeof(int),
2533 		.mode		= 0644,
2534 		.proc_handler	= proc_dointvec_minmax,
2535 		.extra1		= SYSCTL_ZERO,
2536 		.extra2		= SYSCTL_ONE,
2537 	},
2538 	{
2539 		.procname	= "hung_task_check_count",
2540 		.data		= &sysctl_hung_task_check_count,
2541 		.maxlen		= sizeof(int),
2542 		.mode		= 0644,
2543 		.proc_handler	= proc_dointvec_minmax,
2544 		.extra1		= SYSCTL_ZERO,
2545 	},
2546 	{
2547 		.procname	= "hung_task_timeout_secs",
2548 		.data		= &sysctl_hung_task_timeout_secs,
2549 		.maxlen		= sizeof(unsigned long),
2550 		.mode		= 0644,
2551 		.proc_handler	= proc_dohung_task_timeout_secs,
2552 		.extra2		= &hung_task_timeout_max,
2553 	},
2554 	{
2555 		.procname	= "hung_task_check_interval_secs",
2556 		.data		= &sysctl_hung_task_check_interval_secs,
2557 		.maxlen		= sizeof(unsigned long),
2558 		.mode		= 0644,
2559 		.proc_handler	= proc_dohung_task_timeout_secs,
2560 		.extra2		= &hung_task_timeout_max,
2561 	},
2562 	{
2563 		.procname	= "hung_task_warnings",
2564 		.data		= &sysctl_hung_task_warnings,
2565 		.maxlen		= sizeof(int),
2566 		.mode		= 0644,
2567 		.proc_handler	= proc_dointvec_minmax,
2568 		.extra1		= &neg_one,
2569 	},
2570 #endif
2571 #ifdef CONFIG_RT_MUTEXES
2572 	{
2573 		.procname	= "max_lock_depth",
2574 		.data		= &max_lock_depth,
2575 		.maxlen		= sizeof(int),
2576 		.mode		= 0644,
2577 		.proc_handler	= proc_dointvec,
2578 	},
2579 #endif
2580 	{
2581 		.procname	= "poweroff_cmd",
2582 		.data		= &poweroff_cmd,
2583 		.maxlen		= POWEROFF_CMD_PATH_LEN,
2584 		.mode		= 0644,
2585 		.proc_handler	= proc_dostring,
2586 	},
2587 #ifdef CONFIG_KEYS
2588 	{
2589 		.procname	= "keys",
2590 		.mode		= 0555,
2591 		.child		= key_sysctls,
2592 	},
2593 #endif
2594 #ifdef CONFIG_PERF_EVENTS
2595 	/*
2596 	 * User-space scripts rely on the existence of this file
2597 	 * as a feature check for perf_events being enabled.
2598 	 *
2599 	 * So it's an ABI, do not remove!
2600 	 */
2601 	{
2602 		.procname	= "perf_event_paranoid",
2603 		.data		= &sysctl_perf_event_paranoid,
2604 		.maxlen		= sizeof(sysctl_perf_event_paranoid),
2605 		.mode		= 0644,
2606 		.proc_handler	= proc_dointvec,
2607 	},
2608 	{
2609 		.procname	= "perf_event_mlock_kb",
2610 		.data		= &sysctl_perf_event_mlock,
2611 		.maxlen		= sizeof(sysctl_perf_event_mlock),
2612 		.mode		= 0644,
2613 		.proc_handler	= proc_dointvec,
2614 	},
2615 	{
2616 		.procname	= "perf_event_max_sample_rate",
2617 		.data		= &sysctl_perf_event_sample_rate,
2618 		.maxlen		= sizeof(sysctl_perf_event_sample_rate),
2619 		.mode		= 0644,
2620 		.proc_handler	= perf_proc_update_handler,
2621 		.extra1		= SYSCTL_ONE,
2622 	},
2623 	{
2624 		.procname	= "perf_cpu_time_max_percent",
2625 		.data		= &sysctl_perf_cpu_time_max_percent,
2626 		.maxlen		= sizeof(sysctl_perf_cpu_time_max_percent),
2627 		.mode		= 0644,
2628 		.proc_handler	= perf_cpu_time_max_percent_handler,
2629 		.extra1		= SYSCTL_ZERO,
2630 		.extra2		= &one_hundred,
2631 	},
2632 	{
2633 		.procname	= "perf_event_max_stack",
2634 		.data		= &sysctl_perf_event_max_stack,
2635 		.maxlen		= sizeof(sysctl_perf_event_max_stack),
2636 		.mode		= 0644,
2637 		.proc_handler	= perf_event_max_stack_handler,
2638 		.extra1		= SYSCTL_ZERO,
2639 		.extra2		= &six_hundred_forty_kb,
2640 	},
2641 	{
2642 		.procname	= "perf_event_max_contexts_per_stack",
2643 		.data		= &sysctl_perf_event_max_contexts_per_stack,
2644 		.maxlen		= sizeof(sysctl_perf_event_max_contexts_per_stack),
2645 		.mode		= 0644,
2646 		.proc_handler	= perf_event_max_stack_handler,
2647 		.extra1		= SYSCTL_ZERO,
2648 		.extra2		= &one_thousand,
2649 	},
2650 #endif
2651 	{
2652 		.procname	= "panic_on_warn",
2653 		.data		= &panic_on_warn,
2654 		.maxlen		= sizeof(int),
2655 		.mode		= 0644,
2656 		.proc_handler	= proc_dointvec_minmax,
2657 		.extra1		= SYSCTL_ZERO,
2658 		.extra2		= SYSCTL_ONE,
2659 	},
2660 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
2661 	{
2662 		.procname	= "timer_migration",
2663 		.data		= &sysctl_timer_migration,
2664 		.maxlen		= sizeof(unsigned int),
2665 		.mode		= 0644,
2666 		.proc_handler	= timer_migration_handler,
2667 		.extra1		= SYSCTL_ZERO,
2668 		.extra2		= SYSCTL_ONE,
2669 	},
2670 #endif
2671 #ifdef CONFIG_BPF_SYSCALL
2672 	{
2673 		.procname	= "unprivileged_bpf_disabled",
2674 		.data		= &sysctl_unprivileged_bpf_disabled,
2675 		.maxlen		= sizeof(sysctl_unprivileged_bpf_disabled),
2676 		.mode		= 0644,
2677 		.proc_handler	= bpf_unpriv_handler,
2678 		.extra1		= SYSCTL_ZERO,
2679 		.extra2		= &two,
2680 	},
2681 	{
2682 		.procname	= "bpf_stats_enabled",
2683 		.data		= &bpf_stats_enabled_key.key,
2684 		.maxlen		= sizeof(bpf_stats_enabled_key),
2685 		.mode		= 0644,
2686 		.proc_handler	= bpf_stats_handler,
2687 	},
2688 #endif
2689 #if defined(CONFIG_TREE_RCU)
2690 	{
2691 		.procname	= "panic_on_rcu_stall",
2692 		.data		= &sysctl_panic_on_rcu_stall,
2693 		.maxlen		= sizeof(sysctl_panic_on_rcu_stall),
2694 		.mode		= 0644,
2695 		.proc_handler	= proc_dointvec_minmax,
2696 		.extra1		= SYSCTL_ZERO,
2697 		.extra2		= SYSCTL_ONE,
2698 	},
2699 #endif
2700 #if defined(CONFIG_TREE_RCU)
2701 	{
2702 		.procname	= "max_rcu_stall_to_panic",
2703 		.data		= &sysctl_max_rcu_stall_to_panic,
2704 		.maxlen		= sizeof(sysctl_max_rcu_stall_to_panic),
2705 		.mode		= 0644,
2706 		.proc_handler	= proc_dointvec_minmax,
2707 		.extra1		= SYSCTL_ONE,
2708 		.extra2		= SYSCTL_INT_MAX,
2709 	},
2710 #endif
2711 #ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
2712 	{
2713 		.procname	= "stack_erasing",
2714 		.data		= NULL,
2715 		.maxlen		= sizeof(int),
2716 		.mode		= 0600,
2717 		.proc_handler	= stack_erasing_sysctl,
2718 		.extra1		= SYSCTL_ZERO,
2719 		.extra2		= SYSCTL_ONE,
2720 	},
2721 #endif
2722 	{ }
2723 };
2724 
2725 static struct ctl_table vm_table[] = {
2726 	{
2727 		.procname	= "overcommit_memory",
2728 		.data		= &sysctl_overcommit_memory,
2729 		.maxlen		= sizeof(sysctl_overcommit_memory),
2730 		.mode		= 0644,
2731 		.proc_handler	= overcommit_policy_handler,
2732 		.extra1		= SYSCTL_ZERO,
2733 		.extra2		= &two,
2734 	},
2735 	{
2736 		.procname	= "panic_on_oom",
2737 		.data		= &sysctl_panic_on_oom,
2738 		.maxlen		= sizeof(sysctl_panic_on_oom),
2739 		.mode		= 0644,
2740 		.proc_handler	= proc_dointvec_minmax,
2741 		.extra1		= SYSCTL_ZERO,
2742 		.extra2		= &two,
2743 	},
2744 	{
2745 		.procname	= "oom_kill_allocating_task",
2746 		.data		= &sysctl_oom_kill_allocating_task,
2747 		.maxlen		= sizeof(sysctl_oom_kill_allocating_task),
2748 		.mode		= 0644,
2749 		.proc_handler	= proc_dointvec,
2750 	},
2751 	{
2752 		.procname	= "oom_dump_tasks",
2753 		.data		= &sysctl_oom_dump_tasks,
2754 		.maxlen		= sizeof(sysctl_oom_dump_tasks),
2755 		.mode		= 0644,
2756 		.proc_handler	= proc_dointvec,
2757 	},
2758 	{
2759 		.procname	= "overcommit_ratio",
2760 		.data		= &sysctl_overcommit_ratio,
2761 		.maxlen		= sizeof(sysctl_overcommit_ratio),
2762 		.mode		= 0644,
2763 		.proc_handler	= overcommit_ratio_handler,
2764 	},
2765 	{
2766 		.procname	= "overcommit_kbytes",
2767 		.data		= &sysctl_overcommit_kbytes,
2768 		.maxlen		= sizeof(sysctl_overcommit_kbytes),
2769 		.mode		= 0644,
2770 		.proc_handler	= overcommit_kbytes_handler,
2771 	},
2772 	{
2773 		.procname	= "page-cluster",
2774 		.data		= &page_cluster,
2775 		.maxlen		= sizeof(int),
2776 		.mode		= 0644,
2777 		.proc_handler	= proc_dointvec_minmax,
2778 		.extra1		= SYSCTL_ZERO,
2779 	},
2780 	{
2781 		.procname	= "dirty_background_ratio",
2782 		.data		= &dirty_background_ratio,
2783 		.maxlen		= sizeof(dirty_background_ratio),
2784 		.mode		= 0644,
2785 		.proc_handler	= dirty_background_ratio_handler,
2786 		.extra1		= SYSCTL_ZERO,
2787 		.extra2		= &one_hundred,
2788 	},
2789 	{
2790 		.procname	= "dirty_background_bytes",
2791 		.data		= &dirty_background_bytes,
2792 		.maxlen		= sizeof(dirty_background_bytes),
2793 		.mode		= 0644,
2794 		.proc_handler	= dirty_background_bytes_handler,
2795 		.extra1		= &one_ul,
2796 	},
2797 	{
2798 		.procname	= "dirty_ratio",
2799 		.data		= &vm_dirty_ratio,
2800 		.maxlen		= sizeof(vm_dirty_ratio),
2801 		.mode		= 0644,
2802 		.proc_handler	= dirty_ratio_handler,
2803 		.extra1		= SYSCTL_ZERO,
2804 		.extra2		= &one_hundred,
2805 	},
2806 	{
2807 		.procname	= "dirty_bytes",
2808 		.data		= &vm_dirty_bytes,
2809 		.maxlen		= sizeof(vm_dirty_bytes),
2810 		.mode		= 0644,
2811 		.proc_handler	= dirty_bytes_handler,
2812 		.extra1		= &dirty_bytes_min,
2813 	},
2814 	{
2815 		.procname	= "dirty_writeback_centisecs",
2816 		.data		= &dirty_writeback_interval,
2817 		.maxlen		= sizeof(dirty_writeback_interval),
2818 		.mode		= 0644,
2819 		.proc_handler	= dirty_writeback_centisecs_handler,
2820 	},
2821 	{
2822 		.procname	= "dirty_expire_centisecs",
2823 		.data		= &dirty_expire_interval,
2824 		.maxlen		= sizeof(dirty_expire_interval),
2825 		.mode		= 0644,
2826 		.proc_handler	= proc_dointvec_minmax,
2827 		.extra1		= SYSCTL_ZERO,
2828 	},
2829 	{
2830 		.procname	= "dirtytime_expire_seconds",
2831 		.data		= &dirtytime_expire_interval,
2832 		.maxlen		= sizeof(dirtytime_expire_interval),
2833 		.mode		= 0644,
2834 		.proc_handler	= dirtytime_interval_handler,
2835 		.extra1		= SYSCTL_ZERO,
2836 	},
2837 	{
2838 		.procname	= "swappiness",
2839 		.data		= &vm_swappiness,
2840 		.maxlen		= sizeof(vm_swappiness),
2841 		.mode		= 0644,
2842 		.proc_handler	= proc_dointvec_minmax,
2843 		.extra1		= SYSCTL_ZERO,
2844 		.extra2		= &two_hundred,
2845 	},
2846 #ifdef CONFIG_HUGETLB_PAGE
2847 	{
2848 		.procname	= "nr_hugepages",
2849 		.data		= NULL,
2850 		.maxlen		= sizeof(unsigned long),
2851 		.mode		= 0644,
2852 		.proc_handler	= hugetlb_sysctl_handler,
2853 	},
2854 #ifdef CONFIG_NUMA
2855 	{
2856 		.procname       = "nr_hugepages_mempolicy",
2857 		.data           = NULL,
2858 		.maxlen         = sizeof(unsigned long),
2859 		.mode           = 0644,
2860 		.proc_handler   = &hugetlb_mempolicy_sysctl_handler,
2861 	},
2862 	{
2863 		.procname		= "numa_stat",
2864 		.data			= &sysctl_vm_numa_stat,
2865 		.maxlen			= sizeof(int),
2866 		.mode			= 0644,
2867 		.proc_handler	= sysctl_vm_numa_stat_handler,
2868 		.extra1			= SYSCTL_ZERO,
2869 		.extra2			= SYSCTL_ONE,
2870 	},
2871 #endif
2872 	 {
2873 		.procname	= "hugetlb_shm_group",
2874 		.data		= &sysctl_hugetlb_shm_group,
2875 		.maxlen		= sizeof(gid_t),
2876 		.mode		= 0644,
2877 		.proc_handler	= proc_dointvec,
2878 	 },
2879 	{
2880 		.procname	= "nr_overcommit_hugepages",
2881 		.data		= NULL,
2882 		.maxlen		= sizeof(unsigned long),
2883 		.mode		= 0644,
2884 		.proc_handler	= hugetlb_overcommit_handler,
2885 	},
2886 #endif
2887 	{
2888 		.procname	= "lowmem_reserve_ratio",
2889 		.data		= &sysctl_lowmem_reserve_ratio,
2890 		.maxlen		= sizeof(sysctl_lowmem_reserve_ratio),
2891 		.mode		= 0644,
2892 		.proc_handler	= lowmem_reserve_ratio_sysctl_handler,
2893 	},
2894 	{
2895 		.procname	= "drop_caches",
2896 		.data		= &sysctl_drop_caches,
2897 		.maxlen		= sizeof(int),
2898 		.mode		= 0200,
2899 		.proc_handler	= drop_caches_sysctl_handler,
2900 		.extra1		= SYSCTL_ONE,
2901 		.extra2		= &four,
2902 	},
2903 #ifdef CONFIG_COMPACTION
2904 	{
2905 		.procname	= "compact_memory",
2906 		.data		= NULL,
2907 		.maxlen		= sizeof(int),
2908 		.mode		= 0200,
2909 		.proc_handler	= sysctl_compaction_handler,
2910 	},
2911 	{
2912 		.procname	= "compaction_proactiveness",
2913 		.data		= &sysctl_compaction_proactiveness,
2914 		.maxlen		= sizeof(sysctl_compaction_proactiveness),
2915 		.mode		= 0644,
2916 		.proc_handler	= compaction_proactiveness_sysctl_handler,
2917 		.extra1		= SYSCTL_ZERO,
2918 		.extra2		= &one_hundred,
2919 	},
2920 	{
2921 		.procname	= "extfrag_threshold",
2922 		.data		= &sysctl_extfrag_threshold,
2923 		.maxlen		= sizeof(int),
2924 		.mode		= 0644,
2925 		.proc_handler	= proc_dointvec_minmax,
2926 		.extra1		= &min_extfrag_threshold,
2927 		.extra2		= &max_extfrag_threshold,
2928 	},
2929 	{
2930 		.procname	= "compact_unevictable_allowed",
2931 		.data		= &sysctl_compact_unevictable_allowed,
2932 		.maxlen		= sizeof(int),
2933 		.mode		= 0644,
2934 		.proc_handler	= proc_dointvec_minmax_warn_RT_change,
2935 		.extra1		= SYSCTL_ZERO,
2936 		.extra2		= SYSCTL_ONE,
2937 	},
2938 
2939 #endif /* CONFIG_COMPACTION */
2940 	{
2941 		.procname	= "min_free_kbytes",
2942 		.data		= &min_free_kbytes,
2943 		.maxlen		= sizeof(min_free_kbytes),
2944 		.mode		= 0644,
2945 		.proc_handler	= min_free_kbytes_sysctl_handler,
2946 		.extra1		= SYSCTL_ZERO,
2947 	},
2948 	{
2949 		.procname	= "watermark_boost_factor",
2950 		.data		= &watermark_boost_factor,
2951 		.maxlen		= sizeof(watermark_boost_factor),
2952 		.mode		= 0644,
2953 		.proc_handler	= proc_dointvec_minmax,
2954 		.extra1		= SYSCTL_ZERO,
2955 	},
2956 	{
2957 		.procname	= "watermark_scale_factor",
2958 		.data		= &watermark_scale_factor,
2959 		.maxlen		= sizeof(watermark_scale_factor),
2960 		.mode		= 0644,
2961 		.proc_handler	= watermark_scale_factor_sysctl_handler,
2962 		.extra1		= SYSCTL_ONE,
2963 		.extra2		= &one_thousand,
2964 	},
2965 	{
2966 		.procname	= "percpu_pagelist_high_fraction",
2967 		.data		= &percpu_pagelist_high_fraction,
2968 		.maxlen		= sizeof(percpu_pagelist_high_fraction),
2969 		.mode		= 0644,
2970 		.proc_handler	= percpu_pagelist_high_fraction_sysctl_handler,
2971 		.extra1		= SYSCTL_ZERO,
2972 	},
2973 	{
2974 		.procname	= "page_lock_unfairness",
2975 		.data		= &sysctl_page_lock_unfairness,
2976 		.maxlen		= sizeof(sysctl_page_lock_unfairness),
2977 		.mode		= 0644,
2978 		.proc_handler	= proc_dointvec_minmax,
2979 		.extra1		= SYSCTL_ZERO,
2980 	},
2981 #ifdef CONFIG_MMU
2982 	{
2983 		.procname	= "max_map_count",
2984 		.data		= &sysctl_max_map_count,
2985 		.maxlen		= sizeof(sysctl_max_map_count),
2986 		.mode		= 0644,
2987 		.proc_handler	= proc_dointvec_minmax,
2988 		.extra1		= SYSCTL_ZERO,
2989 	},
2990 #else
2991 	{
2992 		.procname	= "nr_trim_pages",
2993 		.data		= &sysctl_nr_trim_pages,
2994 		.maxlen		= sizeof(sysctl_nr_trim_pages),
2995 		.mode		= 0644,
2996 		.proc_handler	= proc_dointvec_minmax,
2997 		.extra1		= SYSCTL_ZERO,
2998 	},
2999 #endif
3000 	{
3001 		.procname	= "laptop_mode",
3002 		.data		= &laptop_mode,
3003 		.maxlen		= sizeof(laptop_mode),
3004 		.mode		= 0644,
3005 		.proc_handler	= proc_dointvec_jiffies,
3006 	},
3007 	{
3008 		.procname	= "vfs_cache_pressure",
3009 		.data		= &sysctl_vfs_cache_pressure,
3010 		.maxlen		= sizeof(sysctl_vfs_cache_pressure),
3011 		.mode		= 0644,
3012 		.proc_handler	= proc_dointvec_minmax,
3013 		.extra1		= SYSCTL_ZERO,
3014 	},
3015 #if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
3016     defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
3017 	{
3018 		.procname	= "legacy_va_layout",
3019 		.data		= &sysctl_legacy_va_layout,
3020 		.maxlen		= sizeof(sysctl_legacy_va_layout),
3021 		.mode		= 0644,
3022 		.proc_handler	= proc_dointvec_minmax,
3023 		.extra1		= SYSCTL_ZERO,
3024 	},
3025 #endif
3026 #ifdef CONFIG_NUMA
3027 	{
3028 		.procname	= "zone_reclaim_mode",
3029 		.data		= &node_reclaim_mode,
3030 		.maxlen		= sizeof(node_reclaim_mode),
3031 		.mode		= 0644,
3032 		.proc_handler	= proc_dointvec_minmax,
3033 		.extra1		= SYSCTL_ZERO,
3034 	},
3035 	{
3036 		.procname	= "min_unmapped_ratio",
3037 		.data		= &sysctl_min_unmapped_ratio,
3038 		.maxlen		= sizeof(sysctl_min_unmapped_ratio),
3039 		.mode		= 0644,
3040 		.proc_handler	= sysctl_min_unmapped_ratio_sysctl_handler,
3041 		.extra1		= SYSCTL_ZERO,
3042 		.extra2		= &one_hundred,
3043 	},
3044 	{
3045 		.procname	= "min_slab_ratio",
3046 		.data		= &sysctl_min_slab_ratio,
3047 		.maxlen		= sizeof(sysctl_min_slab_ratio),
3048 		.mode		= 0644,
3049 		.proc_handler	= sysctl_min_slab_ratio_sysctl_handler,
3050 		.extra1		= SYSCTL_ZERO,
3051 		.extra2		= &one_hundred,
3052 	},
3053 #endif
3054 #ifdef CONFIG_SMP
3055 	{
3056 		.procname	= "stat_interval",
3057 		.data		= &sysctl_stat_interval,
3058 		.maxlen		= sizeof(sysctl_stat_interval),
3059 		.mode		= 0644,
3060 		.proc_handler	= proc_dointvec_jiffies,
3061 	},
3062 	{
3063 		.procname	= "stat_refresh",
3064 		.data		= NULL,
3065 		.maxlen		= 0,
3066 		.mode		= 0600,
3067 		.proc_handler	= vmstat_refresh,
3068 	},
3069 #endif
3070 #ifdef CONFIG_MMU
3071 	{
3072 		.procname	= "mmap_min_addr",
3073 		.data		= &dac_mmap_min_addr,
3074 		.maxlen		= sizeof(unsigned long),
3075 		.mode		= 0644,
3076 		.proc_handler	= mmap_min_addr_handler,
3077 	},
3078 #endif
3079 #ifdef CONFIG_NUMA
3080 	{
3081 		.procname	= "numa_zonelist_order",
3082 		.data		= &numa_zonelist_order,
3083 		.maxlen		= NUMA_ZONELIST_ORDER_LEN,
3084 		.mode		= 0644,
3085 		.proc_handler	= numa_zonelist_order_handler,
3086 	},
3087 #endif
3088 #if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
3089    (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
3090 	{
3091 		.procname	= "vdso_enabled",
3092 #ifdef CONFIG_X86_32
3093 		.data		= &vdso32_enabled,
3094 		.maxlen		= sizeof(vdso32_enabled),
3095 #else
3096 		.data		= &vdso_enabled,
3097 		.maxlen		= sizeof(vdso_enabled),
3098 #endif
3099 		.mode		= 0644,
3100 		.proc_handler	= proc_dointvec,
3101 		.extra1		= SYSCTL_ZERO,
3102 	},
3103 #endif
3104 #ifdef CONFIG_HIGHMEM
3105 	{
3106 		.procname	= "highmem_is_dirtyable",
3107 		.data		= &vm_highmem_is_dirtyable,
3108 		.maxlen		= sizeof(vm_highmem_is_dirtyable),
3109 		.mode		= 0644,
3110 		.proc_handler	= proc_dointvec_minmax,
3111 		.extra1		= SYSCTL_ZERO,
3112 		.extra2		= SYSCTL_ONE,
3113 	},
3114 #endif
3115 #ifdef CONFIG_MEMORY_FAILURE
3116 	{
3117 		.procname	= "memory_failure_early_kill",
3118 		.data		= &sysctl_memory_failure_early_kill,
3119 		.maxlen		= sizeof(sysctl_memory_failure_early_kill),
3120 		.mode		= 0644,
3121 		.proc_handler	= proc_dointvec_minmax,
3122 		.extra1		= SYSCTL_ZERO,
3123 		.extra2		= SYSCTL_ONE,
3124 	},
3125 	{
3126 		.procname	= "memory_failure_recovery",
3127 		.data		= &sysctl_memory_failure_recovery,
3128 		.maxlen		= sizeof(sysctl_memory_failure_recovery),
3129 		.mode		= 0644,
3130 		.proc_handler	= proc_dointvec_minmax,
3131 		.extra1		= SYSCTL_ZERO,
3132 		.extra2		= SYSCTL_ONE,
3133 	},
3134 #endif
3135 	{
3136 		.procname	= "user_reserve_kbytes",
3137 		.data		= &sysctl_user_reserve_kbytes,
3138 		.maxlen		= sizeof(sysctl_user_reserve_kbytes),
3139 		.mode		= 0644,
3140 		.proc_handler	= proc_doulongvec_minmax,
3141 	},
3142 	{
3143 		.procname	= "admin_reserve_kbytes",
3144 		.data		= &sysctl_admin_reserve_kbytes,
3145 		.maxlen		= sizeof(sysctl_admin_reserve_kbytes),
3146 		.mode		= 0644,
3147 		.proc_handler	= proc_doulongvec_minmax,
3148 	},
3149 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_BITS
3150 	{
3151 		.procname	= "mmap_rnd_bits",
3152 		.data		= &mmap_rnd_bits,
3153 		.maxlen		= sizeof(mmap_rnd_bits),
3154 		.mode		= 0600,
3155 		.proc_handler	= proc_dointvec_minmax,
3156 		.extra1		= (void *)&mmap_rnd_bits_min,
3157 		.extra2		= (void *)&mmap_rnd_bits_max,
3158 	},
3159 #endif
3160 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_COMPAT_BITS
3161 	{
3162 		.procname	= "mmap_rnd_compat_bits",
3163 		.data		= &mmap_rnd_compat_bits,
3164 		.maxlen		= sizeof(mmap_rnd_compat_bits),
3165 		.mode		= 0600,
3166 		.proc_handler	= proc_dointvec_minmax,
3167 		.extra1		= (void *)&mmap_rnd_compat_bits_min,
3168 		.extra2		= (void *)&mmap_rnd_compat_bits_max,
3169 	},
3170 #endif
3171 #ifdef CONFIG_USERFAULTFD
3172 	{
3173 		.procname	= "unprivileged_userfaultfd",
3174 		.data		= &sysctl_unprivileged_userfaultfd,
3175 		.maxlen		= sizeof(sysctl_unprivileged_userfaultfd),
3176 		.mode		= 0644,
3177 		.proc_handler	= proc_dointvec_minmax,
3178 		.extra1		= SYSCTL_ZERO,
3179 		.extra2		= SYSCTL_ONE,
3180 	},
3181 #endif
3182 	{ }
3183 };
3184 
3185 static struct ctl_table fs_table[] = {
3186 	{
3187 		.procname	= "inode-nr",
3188 		.data		= &inodes_stat,
3189 		.maxlen		= 2*sizeof(long),
3190 		.mode		= 0444,
3191 		.proc_handler	= proc_nr_inodes,
3192 	},
3193 	{
3194 		.procname	= "inode-state",
3195 		.data		= &inodes_stat,
3196 		.maxlen		= 7*sizeof(long),
3197 		.mode		= 0444,
3198 		.proc_handler	= proc_nr_inodes,
3199 	},
3200 	{
3201 		.procname	= "file-nr",
3202 		.data		= &files_stat,
3203 		.maxlen		= sizeof(files_stat),
3204 		.mode		= 0444,
3205 		.proc_handler	= proc_nr_files,
3206 	},
3207 	{
3208 		.procname	= "file-max",
3209 		.data		= &files_stat.max_files,
3210 		.maxlen		= sizeof(files_stat.max_files),
3211 		.mode		= 0644,
3212 		.proc_handler	= proc_doulongvec_minmax,
3213 		.extra1		= &zero_ul,
3214 		.extra2		= &long_max,
3215 	},
3216 	{
3217 		.procname	= "nr_open",
3218 		.data		= &sysctl_nr_open,
3219 		.maxlen		= sizeof(unsigned int),
3220 		.mode		= 0644,
3221 		.proc_handler	= proc_dointvec_minmax,
3222 		.extra1		= &sysctl_nr_open_min,
3223 		.extra2		= &sysctl_nr_open_max,
3224 	},
3225 	{
3226 		.procname	= "dentry-state",
3227 		.data		= &dentry_stat,
3228 		.maxlen		= 6*sizeof(long),
3229 		.mode		= 0444,
3230 		.proc_handler	= proc_nr_dentry,
3231 	},
3232 	{
3233 		.procname	= "overflowuid",
3234 		.data		= &fs_overflowuid,
3235 		.maxlen		= sizeof(int),
3236 		.mode		= 0644,
3237 		.proc_handler	= proc_dointvec_minmax,
3238 		.extra1		= &minolduid,
3239 		.extra2		= &maxolduid,
3240 	},
3241 	{
3242 		.procname	= "overflowgid",
3243 		.data		= &fs_overflowgid,
3244 		.maxlen		= sizeof(int),
3245 		.mode		= 0644,
3246 		.proc_handler	= proc_dointvec_minmax,
3247 		.extra1		= &minolduid,
3248 		.extra2		= &maxolduid,
3249 	},
3250 #ifdef CONFIG_FILE_LOCKING
3251 	{
3252 		.procname	= "leases-enable",
3253 		.data		= &leases_enable,
3254 		.maxlen		= sizeof(int),
3255 		.mode		= 0644,
3256 		.proc_handler	= proc_dointvec,
3257 	},
3258 #endif
3259 #ifdef CONFIG_DNOTIFY
3260 	{
3261 		.procname	= "dir-notify-enable",
3262 		.data		= &dir_notify_enable,
3263 		.maxlen		= sizeof(int),
3264 		.mode		= 0644,
3265 		.proc_handler	= proc_dointvec,
3266 	},
3267 #endif
3268 #ifdef CONFIG_MMU
3269 #ifdef CONFIG_FILE_LOCKING
3270 	{
3271 		.procname	= "lease-break-time",
3272 		.data		= &lease_break_time,
3273 		.maxlen		= sizeof(int),
3274 		.mode		= 0644,
3275 		.proc_handler	= proc_dointvec,
3276 	},
3277 #endif
3278 #ifdef CONFIG_AIO
3279 	{
3280 		.procname	= "aio-nr",
3281 		.data		= &aio_nr,
3282 		.maxlen		= sizeof(aio_nr),
3283 		.mode		= 0444,
3284 		.proc_handler	= proc_doulongvec_minmax,
3285 	},
3286 	{
3287 		.procname	= "aio-max-nr",
3288 		.data		= &aio_max_nr,
3289 		.maxlen		= sizeof(aio_max_nr),
3290 		.mode		= 0644,
3291 		.proc_handler	= proc_doulongvec_minmax,
3292 	},
3293 #endif /* CONFIG_AIO */
3294 #ifdef CONFIG_INOTIFY_USER
3295 	{
3296 		.procname	= "inotify",
3297 		.mode		= 0555,
3298 		.child		= inotify_table,
3299 	},
3300 #endif
3301 #ifdef CONFIG_FANOTIFY
3302 	{
3303 		.procname	= "fanotify",
3304 		.mode		= 0555,
3305 		.child		= fanotify_table,
3306 	},
3307 #endif
3308 #ifdef CONFIG_EPOLL
3309 	{
3310 		.procname	= "epoll",
3311 		.mode		= 0555,
3312 		.child		= epoll_table,
3313 	},
3314 #endif
3315 #endif
3316 	{
3317 		.procname	= "protected_symlinks",
3318 		.data		= &sysctl_protected_symlinks,
3319 		.maxlen		= sizeof(int),
3320 		.mode		= 0600,
3321 		.proc_handler	= proc_dointvec_minmax,
3322 		.extra1		= SYSCTL_ZERO,
3323 		.extra2		= SYSCTL_ONE,
3324 	},
3325 	{
3326 		.procname	= "protected_hardlinks",
3327 		.data		= &sysctl_protected_hardlinks,
3328 		.maxlen		= sizeof(int),
3329 		.mode		= 0600,
3330 		.proc_handler	= proc_dointvec_minmax,
3331 		.extra1		= SYSCTL_ZERO,
3332 		.extra2		= SYSCTL_ONE,
3333 	},
3334 	{
3335 		.procname	= "protected_fifos",
3336 		.data		= &sysctl_protected_fifos,
3337 		.maxlen		= sizeof(int),
3338 		.mode		= 0600,
3339 		.proc_handler	= proc_dointvec_minmax,
3340 		.extra1		= SYSCTL_ZERO,
3341 		.extra2		= &two,
3342 	},
3343 	{
3344 		.procname	= "protected_regular",
3345 		.data		= &sysctl_protected_regular,
3346 		.maxlen		= sizeof(int),
3347 		.mode		= 0600,
3348 		.proc_handler	= proc_dointvec_minmax,
3349 		.extra1		= SYSCTL_ZERO,
3350 		.extra2		= &two,
3351 	},
3352 	{
3353 		.procname	= "suid_dumpable",
3354 		.data		= &suid_dumpable,
3355 		.maxlen		= sizeof(int),
3356 		.mode		= 0644,
3357 		.proc_handler	= proc_dointvec_minmax_coredump,
3358 		.extra1		= SYSCTL_ZERO,
3359 		.extra2		= &two,
3360 	},
3361 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
3362 	{
3363 		.procname	= "binfmt_misc",
3364 		.mode		= 0555,
3365 		.child		= sysctl_mount_point,
3366 	},
3367 #endif
3368 	{
3369 		.procname	= "pipe-max-size",
3370 		.data		= &pipe_max_size,
3371 		.maxlen		= sizeof(pipe_max_size),
3372 		.mode		= 0644,
3373 		.proc_handler	= proc_dopipe_max_size,
3374 	},
3375 	{
3376 		.procname	= "pipe-user-pages-hard",
3377 		.data		= &pipe_user_pages_hard,
3378 		.maxlen		= sizeof(pipe_user_pages_hard),
3379 		.mode		= 0644,
3380 		.proc_handler	= proc_doulongvec_minmax,
3381 	},
3382 	{
3383 		.procname	= "pipe-user-pages-soft",
3384 		.data		= &pipe_user_pages_soft,
3385 		.maxlen		= sizeof(pipe_user_pages_soft),
3386 		.mode		= 0644,
3387 		.proc_handler	= proc_doulongvec_minmax,
3388 	},
3389 	{
3390 		.procname	= "mount-max",
3391 		.data		= &sysctl_mount_max,
3392 		.maxlen		= sizeof(unsigned int),
3393 		.mode		= 0644,
3394 		.proc_handler	= proc_dointvec_minmax,
3395 		.extra1		= SYSCTL_ONE,
3396 	},
3397 	{ }
3398 };
3399 
3400 static struct ctl_table debug_table[] = {
3401 #ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
3402 	{
3403 		.procname	= "exception-trace",
3404 		.data		= &show_unhandled_signals,
3405 		.maxlen		= sizeof(int),
3406 		.mode		= 0644,
3407 		.proc_handler	= proc_dointvec
3408 	},
3409 #endif
3410 #if defined(CONFIG_OPTPROBES)
3411 	{
3412 		.procname	= "kprobes-optimization",
3413 		.data		= &sysctl_kprobes_optimization,
3414 		.maxlen		= sizeof(int),
3415 		.mode		= 0644,
3416 		.proc_handler	= proc_kprobes_optimization_handler,
3417 		.extra1		= SYSCTL_ZERO,
3418 		.extra2		= SYSCTL_ONE,
3419 	},
3420 #endif
3421 	{ }
3422 };
3423 
3424 static struct ctl_table dev_table[] = {
3425 	{ }
3426 };
3427 
3428 static struct ctl_table sysctl_base_table[] = {
3429 	{
3430 		.procname	= "kernel",
3431 		.mode		= 0555,
3432 		.child		= kern_table,
3433 	},
3434 	{
3435 		.procname	= "vm",
3436 		.mode		= 0555,
3437 		.child		= vm_table,
3438 	},
3439 	{
3440 		.procname	= "fs",
3441 		.mode		= 0555,
3442 		.child		= fs_table,
3443 	},
3444 	{
3445 		.procname	= "debug",
3446 		.mode		= 0555,
3447 		.child		= debug_table,
3448 	},
3449 	{
3450 		.procname	= "dev",
3451 		.mode		= 0555,
3452 		.child		= dev_table,
3453 	},
3454 	{ }
3455 };
3456 
3457 int __init sysctl_init(void)
3458 {
3459 	struct ctl_table_header *hdr;
3460 
3461 	hdr = register_sysctl_table(sysctl_base_table);
3462 	kmemleak_not_leak(hdr);
3463 	return 0;
3464 }
3465 #endif /* CONFIG_SYSCTL */
3466 /*
3467  * No sense putting this after each symbol definition, twice,
3468  * exception granted :-)
3469  */
3470 EXPORT_SYMBOL(proc_dobool);
3471 EXPORT_SYMBOL(proc_dointvec);
3472 EXPORT_SYMBOL(proc_douintvec);
3473 EXPORT_SYMBOL(proc_dointvec_jiffies);
3474 EXPORT_SYMBOL(proc_dointvec_minmax);
3475 EXPORT_SYMBOL_GPL(proc_douintvec_minmax);
3476 EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
3477 EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
3478 EXPORT_SYMBOL(proc_dostring);
3479 EXPORT_SYMBOL(proc_doulongvec_minmax);
3480 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
3481 EXPORT_SYMBOL(proc_do_large_bitmap);
3482