xref: /openbmc/linux/kernel/sysctl.c (revision 8b0adbe3e38dbe5aae9edf6f5159ffdca7cfbdf1)
1 // SPDX-License-Identifier: GPL-2.0-only
2 /*
3  * sysctl.c: General linux system control interface
4  *
5  * Begun 24 March 1995, Stephen Tweedie
6  * Added /proc support, Dec 1995
7  * Added bdflush entry and intvec min/max checking, 2/23/96, Tom Dyas.
8  * Added hooks for /proc/sys/net (minor, minor patch), 96/4/1, Mike Shaver.
9  * Added kernel/java-{interpreter,appletviewer}, 96/5/10, Mike Shaver.
10  * Dynamic registration fixes, Stephen Tweedie.
11  * Added kswapd-interval, ctrl-alt-del, printk stuff, 1/8/97, Chris Horn.
12  * Made sysctl support optional via CONFIG_SYSCTL, 1/10/97, Chris
13  *  Horn.
14  * Added proc_doulongvec_ms_jiffies_minmax, 09/08/99, Carlos H. Bauer.
15  * Added proc_doulongvec_minmax, 09/08/99, Carlos H. Bauer.
16  * Changed linked lists to use list.h instead of lists.h, 02/24/00, Bill
17  *  Wendling.
18  * The list_for_each() macro wasn't appropriate for the sysctl loop.
19  *  Removed it and replaced it with older style, 03/23/00, Bill Wendling
20  */
21 
22 #include <linux/module.h>
23 #include <linux/aio.h>
24 #include <linux/mm.h>
25 #include <linux/swap.h>
26 #include <linux/slab.h>
27 #include <linux/sysctl.h>
28 #include <linux/bitmap.h>
29 #include <linux/signal.h>
30 #include <linux/printk.h>
31 #include <linux/proc_fs.h>
32 #include <linux/security.h>
33 #include <linux/ctype.h>
34 #include <linux/kmemleak.h>
35 #include <linux/fs.h>
36 #include <linux/init.h>
37 #include <linux/kernel.h>
38 #include <linux/kobject.h>
39 #include <linux/net.h>
40 #include <linux/sysrq.h>
41 #include <linux/highuid.h>
42 #include <linux/writeback.h>
43 #include <linux/ratelimit.h>
44 #include <linux/compaction.h>
45 #include <linux/hugetlb.h>
46 #include <linux/initrd.h>
47 #include <linux/key.h>
48 #include <linux/times.h>
49 #include <linux/limits.h>
50 #include <linux/dcache.h>
51 #include <linux/dnotify.h>
52 #include <linux/syscalls.h>
53 #include <linux/vmstat.h>
54 #include <linux/nfs_fs.h>
55 #include <linux/acpi.h>
56 #include <linux/reboot.h>
57 #include <linux/ftrace.h>
58 #include <linux/perf_event.h>
59 #include <linux/kprobes.h>
60 #include <linux/pipe_fs_i.h>
61 #include <linux/oom.h>
62 #include <linux/kmod.h>
63 #include <linux/capability.h>
64 #include <linux/binfmts.h>
65 #include <linux/sched/sysctl.h>
66 #include <linux/sched/coredump.h>
67 #include <linux/kexec.h>
68 #include <linux/bpf.h>
69 #include <linux/mount.h>
70 #include <linux/userfaultfd_k.h>
71 #include <linux/coredump.h>
72 #include <linux/latencytop.h>
73 #include <linux/pid.h>
74 
75 #include "../lib/kstrtox.h"
76 
77 #include <linux/uaccess.h>
78 #include <asm/processor.h>
79 
80 #ifdef CONFIG_X86
81 #include <asm/nmi.h>
82 #include <asm/stacktrace.h>
83 #include <asm/io.h>
84 #endif
85 #ifdef CONFIG_SPARC
86 #include <asm/setup.h>
87 #endif
88 #ifdef CONFIG_BSD_PROCESS_ACCT
89 #include <linux/acct.h>
90 #endif
91 #ifdef CONFIG_RT_MUTEXES
92 #include <linux/rtmutex.h>
93 #endif
94 #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
95 #include <linux/lockdep.h>
96 #endif
97 #ifdef CONFIG_CHR_DEV_SG
98 #include <scsi/sg.h>
99 #endif
100 #ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
101 #include <linux/stackleak.h>
102 #endif
103 #ifdef CONFIG_LOCKUP_DETECTOR
104 #include <linux/nmi.h>
105 #endif
106 
107 #if defined(CONFIG_SYSCTL)
108 
109 /* Constants used for minimum and  maximum */
110 #ifdef CONFIG_LOCKUP_DETECTOR
111 static int sixty = 60;
112 #endif
113 
114 static int __maybe_unused neg_one = -1;
115 static int __maybe_unused two = 2;
116 static int __maybe_unused four = 4;
117 static unsigned long zero_ul;
118 static unsigned long one_ul = 1;
119 static unsigned long long_max = LONG_MAX;
120 static int one_hundred = 100;
121 static int two_hundred = 200;
122 static int one_thousand = 1000;
123 #ifdef CONFIG_PRINTK
124 static int ten_thousand = 10000;
125 #endif
126 #ifdef CONFIG_PERF_EVENTS
127 static int six_hundred_forty_kb = 640 * 1024;
128 #endif
129 
130 /* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
131 static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
132 
133 /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
134 static int maxolduid = 65535;
135 static int minolduid;
136 
137 static int ngroups_max = NGROUPS_MAX;
138 static const int cap_last_cap = CAP_LAST_CAP;
139 
140 /*
141  * This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
142  * and hung_task_check_interval_secs
143  */
144 #ifdef CONFIG_DETECT_HUNG_TASK
145 static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
146 #endif
147 
148 #ifdef CONFIG_INOTIFY_USER
149 #include <linux/inotify.h>
150 #endif
151 
152 #ifdef CONFIG_PROC_SYSCTL
153 
154 /**
155  * enum sysctl_writes_mode - supported sysctl write modes
156  *
157  * @SYSCTL_WRITES_LEGACY: each write syscall must fully contain the sysctl value
158  *	to be written, and multiple writes on the same sysctl file descriptor
159  *	will rewrite the sysctl value, regardless of file position. No warning
160  *	is issued when the initial position is not 0.
161  * @SYSCTL_WRITES_WARN: same as above but warn when the initial file position is
162  *	not 0.
163  * @SYSCTL_WRITES_STRICT: writes to numeric sysctl entries must always be at
164  *	file position 0 and the value must be fully contained in the buffer
165  *	sent to the write syscall. If dealing with strings respect the file
166  *	position, but restrict this to the max length of the buffer, anything
167  *	passed the max length will be ignored. Multiple writes will append
168  *	to the buffer.
169  *
170  * These write modes control how current file position affects the behavior of
171  * updating sysctl values through the proc interface on each write.
172  */
173 enum sysctl_writes_mode {
174 	SYSCTL_WRITES_LEGACY		= -1,
175 	SYSCTL_WRITES_WARN		= 0,
176 	SYSCTL_WRITES_STRICT		= 1,
177 };
178 
179 static enum sysctl_writes_mode sysctl_writes_strict = SYSCTL_WRITES_STRICT;
180 #endif /* CONFIG_PROC_SYSCTL */
181 
182 #if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
183     defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
184 int sysctl_legacy_va_layout;
185 #endif
186 
187 #ifdef CONFIG_SCHED_DEBUG
188 static int min_sched_granularity_ns = 100000;		/* 100 usecs */
189 static int max_sched_granularity_ns = NSEC_PER_SEC;	/* 1 second */
190 static int min_wakeup_granularity_ns;			/* 0 usecs */
191 static int max_wakeup_granularity_ns = NSEC_PER_SEC;	/* 1 second */
192 #ifdef CONFIG_SMP
193 static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
194 static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
195 #endif /* CONFIG_SMP */
196 #endif /* CONFIG_SCHED_DEBUG */
197 
198 #ifdef CONFIG_COMPACTION
199 static int min_extfrag_threshold;
200 static int max_extfrag_threshold = 1000;
201 #endif
202 
203 #endif /* CONFIG_SYSCTL */
204 
205 #if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_SYSCTL)
206 static int bpf_stats_handler(struct ctl_table *table, int write,
207 			     void *buffer, size_t *lenp, loff_t *ppos)
208 {
209 	struct static_key *key = (struct static_key *)table->data;
210 	static int saved_val;
211 	int val, ret;
212 	struct ctl_table tmp = {
213 		.data   = &val,
214 		.maxlen = sizeof(val),
215 		.mode   = table->mode,
216 		.extra1 = SYSCTL_ZERO,
217 		.extra2 = SYSCTL_ONE,
218 	};
219 
220 	if (write && !capable(CAP_SYS_ADMIN))
221 		return -EPERM;
222 
223 	mutex_lock(&bpf_stats_enabled_mutex);
224 	val = saved_val;
225 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
226 	if (write && !ret && val != saved_val) {
227 		if (val)
228 			static_key_slow_inc(key);
229 		else
230 			static_key_slow_dec(key);
231 		saved_val = val;
232 	}
233 	mutex_unlock(&bpf_stats_enabled_mutex);
234 	return ret;
235 }
236 #endif
237 
238 /*
239  * /proc/sys support
240  */
241 
242 #ifdef CONFIG_PROC_SYSCTL
243 
244 static int _proc_do_string(char *data, int maxlen, int write,
245 		char *buffer, size_t *lenp, loff_t *ppos)
246 {
247 	size_t len;
248 	char c, *p;
249 
250 	if (!data || !maxlen || !*lenp) {
251 		*lenp = 0;
252 		return 0;
253 	}
254 
255 	if (write) {
256 		if (sysctl_writes_strict == SYSCTL_WRITES_STRICT) {
257 			/* Only continue writes not past the end of buffer. */
258 			len = strlen(data);
259 			if (len > maxlen - 1)
260 				len = maxlen - 1;
261 
262 			if (*ppos > len)
263 				return 0;
264 			len = *ppos;
265 		} else {
266 			/* Start writing from beginning of buffer. */
267 			len = 0;
268 		}
269 
270 		*ppos += *lenp;
271 		p = buffer;
272 		while ((p - buffer) < *lenp && len < maxlen - 1) {
273 			c = *(p++);
274 			if (c == 0 || c == '\n')
275 				break;
276 			data[len++] = c;
277 		}
278 		data[len] = 0;
279 	} else {
280 		len = strlen(data);
281 		if (len > maxlen)
282 			len = maxlen;
283 
284 		if (*ppos > len) {
285 			*lenp = 0;
286 			return 0;
287 		}
288 
289 		data += *ppos;
290 		len  -= *ppos;
291 
292 		if (len > *lenp)
293 			len = *lenp;
294 		if (len)
295 			memcpy(buffer, data, len);
296 		if (len < *lenp) {
297 			buffer[len] = '\n';
298 			len++;
299 		}
300 		*lenp = len;
301 		*ppos += len;
302 	}
303 	return 0;
304 }
305 
306 static void warn_sysctl_write(struct ctl_table *table)
307 {
308 	pr_warn_once("%s wrote to %s when file position was not 0!\n"
309 		"This will not be supported in the future. To silence this\n"
310 		"warning, set kernel.sysctl_writes_strict = -1\n",
311 		current->comm, table->procname);
312 }
313 
314 /**
315  * proc_first_pos_non_zero_ignore - check if first position is allowed
316  * @ppos: file position
317  * @table: the sysctl table
318  *
319  * Returns true if the first position is non-zero and the sysctl_writes_strict
320  * mode indicates this is not allowed for numeric input types. String proc
321  * handlers can ignore the return value.
322  */
323 static bool proc_first_pos_non_zero_ignore(loff_t *ppos,
324 					   struct ctl_table *table)
325 {
326 	if (!*ppos)
327 		return false;
328 
329 	switch (sysctl_writes_strict) {
330 	case SYSCTL_WRITES_STRICT:
331 		return true;
332 	case SYSCTL_WRITES_WARN:
333 		warn_sysctl_write(table);
334 		return false;
335 	default:
336 		return false;
337 	}
338 }
339 
340 /**
341  * proc_dostring - read a string sysctl
342  * @table: the sysctl table
343  * @write: %TRUE if this is a write to the sysctl file
344  * @buffer: the user buffer
345  * @lenp: the size of the user buffer
346  * @ppos: file position
347  *
348  * Reads/writes a string from/to the user buffer. If the kernel
349  * buffer provided is not large enough to hold the string, the
350  * string is truncated. The copied string is %NULL-terminated.
351  * If the string is being read by the user process, it is copied
352  * and a newline '\n' is added. It is truncated if the buffer is
353  * not large enough.
354  *
355  * Returns 0 on success.
356  */
357 int proc_dostring(struct ctl_table *table, int write,
358 		  void *buffer, size_t *lenp, loff_t *ppos)
359 {
360 	if (write)
361 		proc_first_pos_non_zero_ignore(ppos, table);
362 
363 	return _proc_do_string(table->data, table->maxlen, write, buffer, lenp,
364 			ppos);
365 }
366 
367 static size_t proc_skip_spaces(char **buf)
368 {
369 	size_t ret;
370 	char *tmp = skip_spaces(*buf);
371 	ret = tmp - *buf;
372 	*buf = tmp;
373 	return ret;
374 }
375 
376 static void proc_skip_char(char **buf, size_t *size, const char v)
377 {
378 	while (*size) {
379 		if (**buf != v)
380 			break;
381 		(*size)--;
382 		(*buf)++;
383 	}
384 }
385 
386 /**
387  * strtoul_lenient - parse an ASCII formatted integer from a buffer and only
388  *                   fail on overflow
389  *
390  * @cp: kernel buffer containing the string to parse
391  * @endp: pointer to store the trailing characters
392  * @base: the base to use
393  * @res: where the parsed integer will be stored
394  *
395  * In case of success 0 is returned and @res will contain the parsed integer,
396  * @endp will hold any trailing characters.
397  * This function will fail the parse on overflow. If there wasn't an overflow
398  * the function will defer the decision what characters count as invalid to the
399  * caller.
400  */
401 static int strtoul_lenient(const char *cp, char **endp, unsigned int base,
402 			   unsigned long *res)
403 {
404 	unsigned long long result;
405 	unsigned int rv;
406 
407 	cp = _parse_integer_fixup_radix(cp, &base);
408 	rv = _parse_integer(cp, base, &result);
409 	if ((rv & KSTRTOX_OVERFLOW) || (result != (unsigned long)result))
410 		return -ERANGE;
411 
412 	cp += rv;
413 
414 	if (endp)
415 		*endp = (char *)cp;
416 
417 	*res = (unsigned long)result;
418 	return 0;
419 }
420 
421 #define TMPBUFLEN 22
422 /**
423  * proc_get_long - reads an ASCII formatted integer from a user buffer
424  *
425  * @buf: a kernel buffer
426  * @size: size of the kernel buffer
427  * @val: this is where the number will be stored
428  * @neg: set to %TRUE if number is negative
429  * @perm_tr: a vector which contains the allowed trailers
430  * @perm_tr_len: size of the perm_tr vector
431  * @tr: pointer to store the trailer character
432  *
433  * In case of success %0 is returned and @buf and @size are updated with
434  * the amount of bytes read. If @tr is non-NULL and a trailing
435  * character exists (size is non-zero after returning from this
436  * function), @tr is updated with the trailing character.
437  */
438 static int proc_get_long(char **buf, size_t *size,
439 			  unsigned long *val, bool *neg,
440 			  const char *perm_tr, unsigned perm_tr_len, char *tr)
441 {
442 	int len;
443 	char *p, tmp[TMPBUFLEN];
444 
445 	if (!*size)
446 		return -EINVAL;
447 
448 	len = *size;
449 	if (len > TMPBUFLEN - 1)
450 		len = TMPBUFLEN - 1;
451 
452 	memcpy(tmp, *buf, len);
453 
454 	tmp[len] = 0;
455 	p = tmp;
456 	if (*p == '-' && *size > 1) {
457 		*neg = true;
458 		p++;
459 	} else
460 		*neg = false;
461 	if (!isdigit(*p))
462 		return -EINVAL;
463 
464 	if (strtoul_lenient(p, &p, 0, val))
465 		return -EINVAL;
466 
467 	len = p - tmp;
468 
469 	/* We don't know if the next char is whitespace thus we may accept
470 	 * invalid integers (e.g. 1234...a) or two integers instead of one
471 	 * (e.g. 123...1). So lets not allow such large numbers. */
472 	if (len == TMPBUFLEN - 1)
473 		return -EINVAL;
474 
475 	if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
476 		return -EINVAL;
477 
478 	if (tr && (len < *size))
479 		*tr = *p;
480 
481 	*buf += len;
482 	*size -= len;
483 
484 	return 0;
485 }
486 
487 /**
488  * proc_put_long - converts an integer to a decimal ASCII formatted string
489  *
490  * @buf: the user buffer
491  * @size: the size of the user buffer
492  * @val: the integer to be converted
493  * @neg: sign of the number, %TRUE for negative
494  *
495  * In case of success @buf and @size are updated with the amount of bytes
496  * written.
497  */
498 static void proc_put_long(void **buf, size_t *size, unsigned long val, bool neg)
499 {
500 	int len;
501 	char tmp[TMPBUFLEN], *p = tmp;
502 
503 	sprintf(p, "%s%lu", neg ? "-" : "", val);
504 	len = strlen(tmp);
505 	if (len > *size)
506 		len = *size;
507 	memcpy(*buf, tmp, len);
508 	*size -= len;
509 	*buf += len;
510 }
511 #undef TMPBUFLEN
512 
513 static void proc_put_char(void **buf, size_t *size, char c)
514 {
515 	if (*size) {
516 		char **buffer = (char **)buf;
517 		**buffer = c;
518 
519 		(*size)--;
520 		(*buffer)++;
521 		*buf = *buffer;
522 	}
523 }
524 
525 static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
526 				 int *valp,
527 				 int write, void *data)
528 {
529 	if (write) {
530 		if (*negp) {
531 			if (*lvalp > (unsigned long) INT_MAX + 1)
532 				return -EINVAL;
533 			*valp = -*lvalp;
534 		} else {
535 			if (*lvalp > (unsigned long) INT_MAX)
536 				return -EINVAL;
537 			*valp = *lvalp;
538 		}
539 	} else {
540 		int val = *valp;
541 		if (val < 0) {
542 			*negp = true;
543 			*lvalp = -(unsigned long)val;
544 		} else {
545 			*negp = false;
546 			*lvalp = (unsigned long)val;
547 		}
548 	}
549 	return 0;
550 }
551 
552 static int do_proc_douintvec_conv(unsigned long *lvalp,
553 				  unsigned int *valp,
554 				  int write, void *data)
555 {
556 	if (write) {
557 		if (*lvalp > UINT_MAX)
558 			return -EINVAL;
559 		*valp = *lvalp;
560 	} else {
561 		unsigned int val = *valp;
562 		*lvalp = (unsigned long)val;
563 	}
564 	return 0;
565 }
566 
567 static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
568 
569 static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
570 		  int write, void *buffer,
571 		  size_t *lenp, loff_t *ppos,
572 		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
573 			      int write, void *data),
574 		  void *data)
575 {
576 	int *i, vleft, first = 1, err = 0;
577 	size_t left;
578 	char *p;
579 
580 	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
581 		*lenp = 0;
582 		return 0;
583 	}
584 
585 	i = (int *) tbl_data;
586 	vleft = table->maxlen / sizeof(*i);
587 	left = *lenp;
588 
589 	if (!conv)
590 		conv = do_proc_dointvec_conv;
591 
592 	if (write) {
593 		if (proc_first_pos_non_zero_ignore(ppos, table))
594 			goto out;
595 
596 		if (left > PAGE_SIZE - 1)
597 			left = PAGE_SIZE - 1;
598 		p = buffer;
599 	}
600 
601 	for (; left && vleft--; i++, first=0) {
602 		unsigned long lval;
603 		bool neg;
604 
605 		if (write) {
606 			left -= proc_skip_spaces(&p);
607 
608 			if (!left)
609 				break;
610 			err = proc_get_long(&p, &left, &lval, &neg,
611 					     proc_wspace_sep,
612 					     sizeof(proc_wspace_sep), NULL);
613 			if (err)
614 				break;
615 			if (conv(&neg, &lval, i, 1, data)) {
616 				err = -EINVAL;
617 				break;
618 			}
619 		} else {
620 			if (conv(&neg, &lval, i, 0, data)) {
621 				err = -EINVAL;
622 				break;
623 			}
624 			if (!first)
625 				proc_put_char(&buffer, &left, '\t');
626 			proc_put_long(&buffer, &left, lval, neg);
627 		}
628 	}
629 
630 	if (!write && !first && left && !err)
631 		proc_put_char(&buffer, &left, '\n');
632 	if (write && !err && left)
633 		left -= proc_skip_spaces(&p);
634 	if (write && first)
635 		return err ? : -EINVAL;
636 	*lenp -= left;
637 out:
638 	*ppos += *lenp;
639 	return err;
640 }
641 
642 static int do_proc_dointvec(struct ctl_table *table, int write,
643 		  void *buffer, size_t *lenp, loff_t *ppos,
644 		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
645 			      int write, void *data),
646 		  void *data)
647 {
648 	return __do_proc_dointvec(table->data, table, write,
649 			buffer, lenp, ppos, conv, data);
650 }
651 
652 static int do_proc_douintvec_w(unsigned int *tbl_data,
653 			       struct ctl_table *table,
654 			       void *buffer,
655 			       size_t *lenp, loff_t *ppos,
656 			       int (*conv)(unsigned long *lvalp,
657 					   unsigned int *valp,
658 					   int write, void *data),
659 			       void *data)
660 {
661 	unsigned long lval;
662 	int err = 0;
663 	size_t left;
664 	bool neg;
665 	char *p = buffer;
666 
667 	left = *lenp;
668 
669 	if (proc_first_pos_non_zero_ignore(ppos, table))
670 		goto bail_early;
671 
672 	if (left > PAGE_SIZE - 1)
673 		left = PAGE_SIZE - 1;
674 
675 	left -= proc_skip_spaces(&p);
676 	if (!left) {
677 		err = -EINVAL;
678 		goto out_free;
679 	}
680 
681 	err = proc_get_long(&p, &left, &lval, &neg,
682 			     proc_wspace_sep,
683 			     sizeof(proc_wspace_sep), NULL);
684 	if (err || neg) {
685 		err = -EINVAL;
686 		goto out_free;
687 	}
688 
689 	if (conv(&lval, tbl_data, 1, data)) {
690 		err = -EINVAL;
691 		goto out_free;
692 	}
693 
694 	if (!err && left)
695 		left -= proc_skip_spaces(&p);
696 
697 out_free:
698 	if (err)
699 		return -EINVAL;
700 
701 	return 0;
702 
703 	/* This is in keeping with old __do_proc_dointvec() */
704 bail_early:
705 	*ppos += *lenp;
706 	return err;
707 }
708 
709 static int do_proc_douintvec_r(unsigned int *tbl_data, void *buffer,
710 			       size_t *lenp, loff_t *ppos,
711 			       int (*conv)(unsigned long *lvalp,
712 					   unsigned int *valp,
713 					   int write, void *data),
714 			       void *data)
715 {
716 	unsigned long lval;
717 	int err = 0;
718 	size_t left;
719 
720 	left = *lenp;
721 
722 	if (conv(&lval, tbl_data, 0, data)) {
723 		err = -EINVAL;
724 		goto out;
725 	}
726 
727 	proc_put_long(&buffer, &left, lval, false);
728 	if (!left)
729 		goto out;
730 
731 	proc_put_char(&buffer, &left, '\n');
732 
733 out:
734 	*lenp -= left;
735 	*ppos += *lenp;
736 
737 	return err;
738 }
739 
740 static int __do_proc_douintvec(void *tbl_data, struct ctl_table *table,
741 			       int write, void *buffer,
742 			       size_t *lenp, loff_t *ppos,
743 			       int (*conv)(unsigned long *lvalp,
744 					   unsigned int *valp,
745 					   int write, void *data),
746 			       void *data)
747 {
748 	unsigned int *i, vleft;
749 
750 	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
751 		*lenp = 0;
752 		return 0;
753 	}
754 
755 	i = (unsigned int *) tbl_data;
756 	vleft = table->maxlen / sizeof(*i);
757 
758 	/*
759 	 * Arrays are not supported, keep this simple. *Do not* add
760 	 * support for them.
761 	 */
762 	if (vleft != 1) {
763 		*lenp = 0;
764 		return -EINVAL;
765 	}
766 
767 	if (!conv)
768 		conv = do_proc_douintvec_conv;
769 
770 	if (write)
771 		return do_proc_douintvec_w(i, table, buffer, lenp, ppos,
772 					   conv, data);
773 	return do_proc_douintvec_r(i, buffer, lenp, ppos, conv, data);
774 }
775 
776 static int do_proc_douintvec(struct ctl_table *table, int write,
777 			     void *buffer, size_t *lenp, loff_t *ppos,
778 			     int (*conv)(unsigned long *lvalp,
779 					 unsigned int *valp,
780 					 int write, void *data),
781 			     void *data)
782 {
783 	return __do_proc_douintvec(table->data, table, write,
784 				   buffer, lenp, ppos, conv, data);
785 }
786 
787 /**
788  * proc_dointvec - read a vector of integers
789  * @table: the sysctl table
790  * @write: %TRUE if this is a write to the sysctl file
791  * @buffer: the user buffer
792  * @lenp: the size of the user buffer
793  * @ppos: file position
794  *
795  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
796  * values from/to the user buffer, treated as an ASCII string.
797  *
798  * Returns 0 on success.
799  */
800 int proc_dointvec(struct ctl_table *table, int write, void *buffer,
801 		  size_t *lenp, loff_t *ppos)
802 {
803 	return do_proc_dointvec(table, write, buffer, lenp, ppos, NULL, NULL);
804 }
805 
806 #ifdef CONFIG_COMPACTION
807 static int proc_dointvec_minmax_warn_RT_change(struct ctl_table *table,
808 		int write, void *buffer, size_t *lenp, loff_t *ppos)
809 {
810 	int ret, old;
811 
812 	if (!IS_ENABLED(CONFIG_PREEMPT_RT) || !write)
813 		return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
814 
815 	old = *(int *)table->data;
816 	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
817 	if (ret)
818 		return ret;
819 	if (old != *(int *)table->data)
820 		pr_warn_once("sysctl attribute %s changed by %s[%d]\n",
821 			     table->procname, current->comm,
822 			     task_pid_nr(current));
823 	return ret;
824 }
825 #endif
826 
827 /**
828  * proc_douintvec - read a vector of unsigned integers
829  * @table: the sysctl table
830  * @write: %TRUE if this is a write to the sysctl file
831  * @buffer: the user buffer
832  * @lenp: the size of the user buffer
833  * @ppos: file position
834  *
835  * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
836  * values from/to the user buffer, treated as an ASCII string.
837  *
838  * Returns 0 on success.
839  */
840 int proc_douintvec(struct ctl_table *table, int write, void *buffer,
841 		size_t *lenp, loff_t *ppos)
842 {
843 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
844 				 do_proc_douintvec_conv, NULL);
845 }
846 
847 /*
848  * Taint values can only be increased
849  * This means we can safely use a temporary.
850  */
851 static int proc_taint(struct ctl_table *table, int write,
852 			       void *buffer, size_t *lenp, loff_t *ppos)
853 {
854 	struct ctl_table t;
855 	unsigned long tmptaint = get_taint();
856 	int err;
857 
858 	if (write && !capable(CAP_SYS_ADMIN))
859 		return -EPERM;
860 
861 	t = *table;
862 	t.data = &tmptaint;
863 	err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
864 	if (err < 0)
865 		return err;
866 
867 	if (write) {
868 		int i;
869 
870 		/*
871 		 * If we are relying on panic_on_taint not producing
872 		 * false positives due to userspace input, bail out
873 		 * before setting the requested taint flags.
874 		 */
875 		if (panic_on_taint_nousertaint && (tmptaint & panic_on_taint))
876 			return -EINVAL;
877 
878 		/*
879 		 * Poor man's atomic or. Not worth adding a primitive
880 		 * to everyone's atomic.h for this
881 		 */
882 		for (i = 0; i < TAINT_FLAGS_COUNT; i++)
883 			if ((1UL << i) & tmptaint)
884 				add_taint(i, LOCKDEP_STILL_OK);
885 	}
886 
887 	return err;
888 }
889 
890 #ifdef CONFIG_PRINTK
891 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
892 				void *buffer, size_t *lenp, loff_t *ppos)
893 {
894 	if (write && !capable(CAP_SYS_ADMIN))
895 		return -EPERM;
896 
897 	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
898 }
899 #endif
900 
901 /**
902  * struct do_proc_dointvec_minmax_conv_param - proc_dointvec_minmax() range checking structure
903  * @min: pointer to minimum allowable value
904  * @max: pointer to maximum allowable value
905  *
906  * The do_proc_dointvec_minmax_conv_param structure provides the
907  * minimum and maximum values for doing range checking for those sysctl
908  * parameters that use the proc_dointvec_minmax() handler.
909  */
910 struct do_proc_dointvec_minmax_conv_param {
911 	int *min;
912 	int *max;
913 };
914 
915 static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
916 					int *valp,
917 					int write, void *data)
918 {
919 	int tmp, ret;
920 	struct do_proc_dointvec_minmax_conv_param *param = data;
921 	/*
922 	 * If writing, first do so via a temporary local int so we can
923 	 * bounds-check it before touching *valp.
924 	 */
925 	int *ip = write ? &tmp : valp;
926 
927 	ret = do_proc_dointvec_conv(negp, lvalp, ip, write, data);
928 	if (ret)
929 		return ret;
930 
931 	if (write) {
932 		if ((param->min && *param->min > tmp) ||
933 		    (param->max && *param->max < tmp))
934 			return -EINVAL;
935 		*valp = tmp;
936 	}
937 
938 	return 0;
939 }
940 
941 /**
942  * proc_dointvec_minmax - read a vector of integers with min/max values
943  * @table: the sysctl table
944  * @write: %TRUE if this is a write to the sysctl file
945  * @buffer: the user buffer
946  * @lenp: the size of the user buffer
947  * @ppos: file position
948  *
949  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
950  * values from/to the user buffer, treated as an ASCII string.
951  *
952  * This routine will ensure the values are within the range specified by
953  * table->extra1 (min) and table->extra2 (max).
954  *
955  * Returns 0 on success or -EINVAL on write when the range check fails.
956  */
957 int proc_dointvec_minmax(struct ctl_table *table, int write,
958 		  void *buffer, size_t *lenp, loff_t *ppos)
959 {
960 	struct do_proc_dointvec_minmax_conv_param param = {
961 		.min = (int *) table->extra1,
962 		.max = (int *) table->extra2,
963 	};
964 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
965 				do_proc_dointvec_minmax_conv, &param);
966 }
967 
968 /**
969  * struct do_proc_douintvec_minmax_conv_param - proc_douintvec_minmax() range checking structure
970  * @min: pointer to minimum allowable value
971  * @max: pointer to maximum allowable value
972  *
973  * The do_proc_douintvec_minmax_conv_param structure provides the
974  * minimum and maximum values for doing range checking for those sysctl
975  * parameters that use the proc_douintvec_minmax() handler.
976  */
977 struct do_proc_douintvec_minmax_conv_param {
978 	unsigned int *min;
979 	unsigned int *max;
980 };
981 
982 static int do_proc_douintvec_minmax_conv(unsigned long *lvalp,
983 					 unsigned int *valp,
984 					 int write, void *data)
985 {
986 	int ret;
987 	unsigned int tmp;
988 	struct do_proc_douintvec_minmax_conv_param *param = data;
989 	/* write via temporary local uint for bounds-checking */
990 	unsigned int *up = write ? &tmp : valp;
991 
992 	ret = do_proc_douintvec_conv(lvalp, up, write, data);
993 	if (ret)
994 		return ret;
995 
996 	if (write) {
997 		if ((param->min && *param->min > tmp) ||
998 		    (param->max && *param->max < tmp))
999 			return -ERANGE;
1000 
1001 		*valp = tmp;
1002 	}
1003 
1004 	return 0;
1005 }
1006 
1007 /**
1008  * proc_douintvec_minmax - read a vector of unsigned ints with min/max values
1009  * @table: the sysctl table
1010  * @write: %TRUE if this is a write to the sysctl file
1011  * @buffer: the user buffer
1012  * @lenp: the size of the user buffer
1013  * @ppos: file position
1014  *
1015  * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
1016  * values from/to the user buffer, treated as an ASCII string. Negative
1017  * strings are not allowed.
1018  *
1019  * This routine will ensure the values are within the range specified by
1020  * table->extra1 (min) and table->extra2 (max). There is a final sanity
1021  * check for UINT_MAX to avoid having to support wrap around uses from
1022  * userspace.
1023  *
1024  * Returns 0 on success or -ERANGE on write when the range check fails.
1025  */
1026 int proc_douintvec_minmax(struct ctl_table *table, int write,
1027 			  void *buffer, size_t *lenp, loff_t *ppos)
1028 {
1029 	struct do_proc_douintvec_minmax_conv_param param = {
1030 		.min = (unsigned int *) table->extra1,
1031 		.max = (unsigned int *) table->extra2,
1032 	};
1033 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1034 				 do_proc_douintvec_minmax_conv, &param);
1035 }
1036 
1037 /**
1038  * proc_dou8vec_minmax - read a vector of unsigned chars with min/max values
1039  * @table: the sysctl table
1040  * @write: %TRUE if this is a write to the sysctl file
1041  * @buffer: the user buffer
1042  * @lenp: the size of the user buffer
1043  * @ppos: file position
1044  *
1045  * Reads/writes up to table->maxlen/sizeof(u8) unsigned chars
1046  * values from/to the user buffer, treated as an ASCII string. Negative
1047  * strings are not allowed.
1048  *
1049  * This routine will ensure the values are within the range specified by
1050  * table->extra1 (min) and table->extra2 (max).
1051  *
1052  * Returns 0 on success or an error on write when the range check fails.
1053  */
1054 int proc_dou8vec_minmax(struct ctl_table *table, int write,
1055 			void *buffer, size_t *lenp, loff_t *ppos)
1056 {
1057 	struct ctl_table tmp;
1058 	unsigned int min = 0, max = 255U, val;
1059 	u8 *data = table->data;
1060 	struct do_proc_douintvec_minmax_conv_param param = {
1061 		.min = &min,
1062 		.max = &max,
1063 	};
1064 	int res;
1065 
1066 	/* Do not support arrays yet. */
1067 	if (table->maxlen != sizeof(u8))
1068 		return -EINVAL;
1069 
1070 	if (table->extra1) {
1071 		min = *(unsigned int *) table->extra1;
1072 		if (min > 255U)
1073 			return -EINVAL;
1074 	}
1075 	if (table->extra2) {
1076 		max = *(unsigned int *) table->extra2;
1077 		if (max > 255U)
1078 			return -EINVAL;
1079 	}
1080 
1081 	tmp = *table;
1082 
1083 	tmp.maxlen = sizeof(val);
1084 	tmp.data = &val;
1085 	val = *data;
1086 	res = do_proc_douintvec(&tmp, write, buffer, lenp, ppos,
1087 				do_proc_douintvec_minmax_conv, &param);
1088 	if (res)
1089 		return res;
1090 	if (write)
1091 		*data = val;
1092 	return 0;
1093 }
1094 EXPORT_SYMBOL_GPL(proc_dou8vec_minmax);
1095 
1096 static int do_proc_dopipe_max_size_conv(unsigned long *lvalp,
1097 					unsigned int *valp,
1098 					int write, void *data)
1099 {
1100 	if (write) {
1101 		unsigned int val;
1102 
1103 		val = round_pipe_size(*lvalp);
1104 		if (val == 0)
1105 			return -EINVAL;
1106 
1107 		*valp = val;
1108 	} else {
1109 		unsigned int val = *valp;
1110 		*lvalp = (unsigned long) val;
1111 	}
1112 
1113 	return 0;
1114 }
1115 
1116 static int proc_dopipe_max_size(struct ctl_table *table, int write,
1117 				void *buffer, size_t *lenp, loff_t *ppos)
1118 {
1119 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1120 				 do_proc_dopipe_max_size_conv, NULL);
1121 }
1122 
1123 static void validate_coredump_safety(void)
1124 {
1125 #ifdef CONFIG_COREDUMP
1126 	if (suid_dumpable == SUID_DUMP_ROOT &&
1127 	    core_pattern[0] != '/' && core_pattern[0] != '|') {
1128 		printk(KERN_WARNING
1129 "Unsafe core_pattern used with fs.suid_dumpable=2.\n"
1130 "Pipe handler or fully qualified core dump path required.\n"
1131 "Set kernel.core_pattern before fs.suid_dumpable.\n"
1132 		);
1133 	}
1134 #endif
1135 }
1136 
1137 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
1138 		void *buffer, size_t *lenp, loff_t *ppos)
1139 {
1140 	int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
1141 	if (!error)
1142 		validate_coredump_safety();
1143 	return error;
1144 }
1145 
1146 #ifdef CONFIG_COREDUMP
1147 static int proc_dostring_coredump(struct ctl_table *table, int write,
1148 		  void *buffer, size_t *lenp, loff_t *ppos)
1149 {
1150 	int error = proc_dostring(table, write, buffer, lenp, ppos);
1151 	if (!error)
1152 		validate_coredump_safety();
1153 	return error;
1154 }
1155 #endif
1156 
1157 #ifdef CONFIG_MAGIC_SYSRQ
1158 static int sysrq_sysctl_handler(struct ctl_table *table, int write,
1159 				void *buffer, size_t *lenp, loff_t *ppos)
1160 {
1161 	int tmp, ret;
1162 
1163 	tmp = sysrq_mask();
1164 
1165 	ret = __do_proc_dointvec(&tmp, table, write, buffer,
1166 			       lenp, ppos, NULL, NULL);
1167 	if (ret || !write)
1168 		return ret;
1169 
1170 	if (write)
1171 		sysrq_toggle_support(tmp);
1172 
1173 	return 0;
1174 }
1175 #endif
1176 
1177 static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table,
1178 		int write, void *buffer, size_t *lenp, loff_t *ppos,
1179 		unsigned long convmul, unsigned long convdiv)
1180 {
1181 	unsigned long *i, *min, *max;
1182 	int vleft, first = 1, err = 0;
1183 	size_t left;
1184 	char *p;
1185 
1186 	if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
1187 		*lenp = 0;
1188 		return 0;
1189 	}
1190 
1191 	i = (unsigned long *) data;
1192 	min = (unsigned long *) table->extra1;
1193 	max = (unsigned long *) table->extra2;
1194 	vleft = table->maxlen / sizeof(unsigned long);
1195 	left = *lenp;
1196 
1197 	if (write) {
1198 		if (proc_first_pos_non_zero_ignore(ppos, table))
1199 			goto out;
1200 
1201 		if (left > PAGE_SIZE - 1)
1202 			left = PAGE_SIZE - 1;
1203 		p = buffer;
1204 	}
1205 
1206 	for (; left && vleft--; i++, first = 0) {
1207 		unsigned long val;
1208 
1209 		if (write) {
1210 			bool neg;
1211 
1212 			left -= proc_skip_spaces(&p);
1213 			if (!left)
1214 				break;
1215 
1216 			err = proc_get_long(&p, &left, &val, &neg,
1217 					     proc_wspace_sep,
1218 					     sizeof(proc_wspace_sep), NULL);
1219 			if (err)
1220 				break;
1221 			if (neg)
1222 				continue;
1223 			val = convmul * val / convdiv;
1224 			if ((min && val < *min) || (max && val > *max)) {
1225 				err = -EINVAL;
1226 				break;
1227 			}
1228 			*i = val;
1229 		} else {
1230 			val = convdiv * (*i) / convmul;
1231 			if (!first)
1232 				proc_put_char(&buffer, &left, '\t');
1233 			proc_put_long(&buffer, &left, val, false);
1234 		}
1235 	}
1236 
1237 	if (!write && !first && left && !err)
1238 		proc_put_char(&buffer, &left, '\n');
1239 	if (write && !err)
1240 		left -= proc_skip_spaces(&p);
1241 	if (write && first)
1242 		return err ? : -EINVAL;
1243 	*lenp -= left;
1244 out:
1245 	*ppos += *lenp;
1246 	return err;
1247 }
1248 
1249 static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
1250 		void *buffer, size_t *lenp, loff_t *ppos, unsigned long convmul,
1251 		unsigned long convdiv)
1252 {
1253 	return __do_proc_doulongvec_minmax(table->data, table, write,
1254 			buffer, lenp, ppos, convmul, convdiv);
1255 }
1256 
1257 /**
1258  * proc_doulongvec_minmax - read a vector of long integers with min/max values
1259  * @table: the sysctl table
1260  * @write: %TRUE if this is a write to the sysctl file
1261  * @buffer: the user buffer
1262  * @lenp: the size of the user buffer
1263  * @ppos: file position
1264  *
1265  * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1266  * values from/to the user buffer, treated as an ASCII string.
1267  *
1268  * This routine will ensure the values are within the range specified by
1269  * table->extra1 (min) and table->extra2 (max).
1270  *
1271  * Returns 0 on success.
1272  */
1273 int proc_doulongvec_minmax(struct ctl_table *table, int write,
1274 			   void *buffer, size_t *lenp, loff_t *ppos)
1275 {
1276     return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
1277 }
1278 
1279 /**
1280  * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
1281  * @table: the sysctl table
1282  * @write: %TRUE if this is a write to the sysctl file
1283  * @buffer: the user buffer
1284  * @lenp: the size of the user buffer
1285  * @ppos: file position
1286  *
1287  * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1288  * values from/to the user buffer, treated as an ASCII string. The values
1289  * are treated as milliseconds, and converted to jiffies when they are stored.
1290  *
1291  * This routine will ensure the values are within the range specified by
1292  * table->extra1 (min) and table->extra2 (max).
1293  *
1294  * Returns 0 on success.
1295  */
1296 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1297 				      void *buffer, size_t *lenp, loff_t *ppos)
1298 {
1299     return do_proc_doulongvec_minmax(table, write, buffer,
1300 				     lenp, ppos, HZ, 1000l);
1301 }
1302 
1303 
1304 static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
1305 					 int *valp,
1306 					 int write, void *data)
1307 {
1308 	if (write) {
1309 		if (*lvalp > INT_MAX / HZ)
1310 			return 1;
1311 		*valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
1312 	} else {
1313 		int val = *valp;
1314 		unsigned long lval;
1315 		if (val < 0) {
1316 			*negp = true;
1317 			lval = -(unsigned long)val;
1318 		} else {
1319 			*negp = false;
1320 			lval = (unsigned long)val;
1321 		}
1322 		*lvalp = lval / HZ;
1323 	}
1324 	return 0;
1325 }
1326 
1327 static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
1328 						int *valp,
1329 						int write, void *data)
1330 {
1331 	if (write) {
1332 		if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
1333 			return 1;
1334 		*valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
1335 	} else {
1336 		int val = *valp;
1337 		unsigned long lval;
1338 		if (val < 0) {
1339 			*negp = true;
1340 			lval = -(unsigned long)val;
1341 		} else {
1342 			*negp = false;
1343 			lval = (unsigned long)val;
1344 		}
1345 		*lvalp = jiffies_to_clock_t(lval);
1346 	}
1347 	return 0;
1348 }
1349 
1350 static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
1351 					    int *valp,
1352 					    int write, void *data)
1353 {
1354 	if (write) {
1355 		unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
1356 
1357 		if (jif > INT_MAX)
1358 			return 1;
1359 		*valp = (int)jif;
1360 	} else {
1361 		int val = *valp;
1362 		unsigned long lval;
1363 		if (val < 0) {
1364 			*negp = true;
1365 			lval = -(unsigned long)val;
1366 		} else {
1367 			*negp = false;
1368 			lval = (unsigned long)val;
1369 		}
1370 		*lvalp = jiffies_to_msecs(lval);
1371 	}
1372 	return 0;
1373 }
1374 
1375 /**
1376  * proc_dointvec_jiffies - read a vector of integers as seconds
1377  * @table: the sysctl table
1378  * @write: %TRUE if this is a write to the sysctl file
1379  * @buffer: the user buffer
1380  * @lenp: the size of the user buffer
1381  * @ppos: file position
1382  *
1383  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1384  * values from/to the user buffer, treated as an ASCII string.
1385  * The values read are assumed to be in seconds, and are converted into
1386  * jiffies.
1387  *
1388  * Returns 0 on success.
1389  */
1390 int proc_dointvec_jiffies(struct ctl_table *table, int write,
1391 			  void *buffer, size_t *lenp, loff_t *ppos)
1392 {
1393     return do_proc_dointvec(table,write,buffer,lenp,ppos,
1394 		    	    do_proc_dointvec_jiffies_conv,NULL);
1395 }
1396 
1397 /**
1398  * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
1399  * @table: the sysctl table
1400  * @write: %TRUE if this is a write to the sysctl file
1401  * @buffer: the user buffer
1402  * @lenp: the size of the user buffer
1403  * @ppos: pointer to the file position
1404  *
1405  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1406  * values from/to the user buffer, treated as an ASCII string.
1407  * The values read are assumed to be in 1/USER_HZ seconds, and
1408  * are converted into jiffies.
1409  *
1410  * Returns 0 on success.
1411  */
1412 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1413 				 void *buffer, size_t *lenp, loff_t *ppos)
1414 {
1415     return do_proc_dointvec(table,write,buffer,lenp,ppos,
1416 		    	    do_proc_dointvec_userhz_jiffies_conv,NULL);
1417 }
1418 
1419 /**
1420  * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
1421  * @table: the sysctl table
1422  * @write: %TRUE if this is a write to the sysctl file
1423  * @buffer: the user buffer
1424  * @lenp: the size of the user buffer
1425  * @ppos: file position
1426  * @ppos: the current position in the file
1427  *
1428  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1429  * values from/to the user buffer, treated as an ASCII string.
1430  * The values read are assumed to be in 1/1000 seconds, and
1431  * are converted into jiffies.
1432  *
1433  * Returns 0 on success.
1434  */
1435 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write, void *buffer,
1436 		size_t *lenp, loff_t *ppos)
1437 {
1438 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
1439 				do_proc_dointvec_ms_jiffies_conv, NULL);
1440 }
1441 
1442 static int proc_do_cad_pid(struct ctl_table *table, int write, void *buffer,
1443 		size_t *lenp, loff_t *ppos)
1444 {
1445 	struct pid *new_pid;
1446 	pid_t tmp;
1447 	int r;
1448 
1449 	tmp = pid_vnr(cad_pid);
1450 
1451 	r = __do_proc_dointvec(&tmp, table, write, buffer,
1452 			       lenp, ppos, NULL, NULL);
1453 	if (r || !write)
1454 		return r;
1455 
1456 	new_pid = find_get_pid(tmp);
1457 	if (!new_pid)
1458 		return -ESRCH;
1459 
1460 	put_pid(xchg(&cad_pid, new_pid));
1461 	return 0;
1462 }
1463 
1464 /**
1465  * proc_do_large_bitmap - read/write from/to a large bitmap
1466  * @table: the sysctl table
1467  * @write: %TRUE if this is a write to the sysctl file
1468  * @buffer: the user buffer
1469  * @lenp: the size of the user buffer
1470  * @ppos: file position
1471  *
1472  * The bitmap is stored at table->data and the bitmap length (in bits)
1473  * in table->maxlen.
1474  *
1475  * We use a range comma separated format (e.g. 1,3-4,10-10) so that
1476  * large bitmaps may be represented in a compact manner. Writing into
1477  * the file will clear the bitmap then update it with the given input.
1478  *
1479  * Returns 0 on success.
1480  */
1481 int proc_do_large_bitmap(struct ctl_table *table, int write,
1482 			 void *buffer, size_t *lenp, loff_t *ppos)
1483 {
1484 	int err = 0;
1485 	bool first = 1;
1486 	size_t left = *lenp;
1487 	unsigned long bitmap_len = table->maxlen;
1488 	unsigned long *bitmap = *(unsigned long **) table->data;
1489 	unsigned long *tmp_bitmap = NULL;
1490 	char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
1491 
1492 	if (!bitmap || !bitmap_len || !left || (*ppos && !write)) {
1493 		*lenp = 0;
1494 		return 0;
1495 	}
1496 
1497 	if (write) {
1498 		char *p = buffer;
1499 		size_t skipped = 0;
1500 
1501 		if (left > PAGE_SIZE - 1) {
1502 			left = PAGE_SIZE - 1;
1503 			/* How much of the buffer we'll skip this pass */
1504 			skipped = *lenp - left;
1505 		}
1506 
1507 		tmp_bitmap = bitmap_zalloc(bitmap_len, GFP_KERNEL);
1508 		if (!tmp_bitmap)
1509 			return -ENOMEM;
1510 		proc_skip_char(&p, &left, '\n');
1511 		while (!err && left) {
1512 			unsigned long val_a, val_b;
1513 			bool neg;
1514 			size_t saved_left;
1515 
1516 			/* In case we stop parsing mid-number, we can reset */
1517 			saved_left = left;
1518 			err = proc_get_long(&p, &left, &val_a, &neg, tr_a,
1519 					     sizeof(tr_a), &c);
1520 			/*
1521 			 * If we consumed the entirety of a truncated buffer or
1522 			 * only one char is left (may be a "-"), then stop here,
1523 			 * reset, & come back for more.
1524 			 */
1525 			if ((left <= 1) && skipped) {
1526 				left = saved_left;
1527 				break;
1528 			}
1529 
1530 			if (err)
1531 				break;
1532 			if (val_a >= bitmap_len || neg) {
1533 				err = -EINVAL;
1534 				break;
1535 			}
1536 
1537 			val_b = val_a;
1538 			if (left) {
1539 				p++;
1540 				left--;
1541 			}
1542 
1543 			if (c == '-') {
1544 				err = proc_get_long(&p, &left, &val_b,
1545 						     &neg, tr_b, sizeof(tr_b),
1546 						     &c);
1547 				/*
1548 				 * If we consumed all of a truncated buffer or
1549 				 * then stop here, reset, & come back for more.
1550 				 */
1551 				if (!left && skipped) {
1552 					left = saved_left;
1553 					break;
1554 				}
1555 
1556 				if (err)
1557 					break;
1558 				if (val_b >= bitmap_len || neg ||
1559 				    val_a > val_b) {
1560 					err = -EINVAL;
1561 					break;
1562 				}
1563 				if (left) {
1564 					p++;
1565 					left--;
1566 				}
1567 			}
1568 
1569 			bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
1570 			first = 0;
1571 			proc_skip_char(&p, &left, '\n');
1572 		}
1573 		left += skipped;
1574 	} else {
1575 		unsigned long bit_a, bit_b = 0;
1576 
1577 		while (left) {
1578 			bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
1579 			if (bit_a >= bitmap_len)
1580 				break;
1581 			bit_b = find_next_zero_bit(bitmap, bitmap_len,
1582 						   bit_a + 1) - 1;
1583 
1584 			if (!first)
1585 				proc_put_char(&buffer, &left, ',');
1586 			proc_put_long(&buffer, &left, bit_a, false);
1587 			if (bit_a != bit_b) {
1588 				proc_put_char(&buffer, &left, '-');
1589 				proc_put_long(&buffer, &left, bit_b, false);
1590 			}
1591 
1592 			first = 0; bit_b++;
1593 		}
1594 		proc_put_char(&buffer, &left, '\n');
1595 	}
1596 
1597 	if (!err) {
1598 		if (write) {
1599 			if (*ppos)
1600 				bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
1601 			else
1602 				bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
1603 		}
1604 		*lenp -= left;
1605 		*ppos += *lenp;
1606 	}
1607 
1608 	bitmap_free(tmp_bitmap);
1609 	return err;
1610 }
1611 
1612 #else /* CONFIG_PROC_SYSCTL */
1613 
1614 int proc_dostring(struct ctl_table *table, int write,
1615 		  void *buffer, size_t *lenp, loff_t *ppos)
1616 {
1617 	return -ENOSYS;
1618 }
1619 
1620 int proc_dointvec(struct ctl_table *table, int write,
1621 		  void *buffer, size_t *lenp, loff_t *ppos)
1622 {
1623 	return -ENOSYS;
1624 }
1625 
1626 int proc_douintvec(struct ctl_table *table, int write,
1627 		  void *buffer, size_t *lenp, loff_t *ppos)
1628 {
1629 	return -ENOSYS;
1630 }
1631 
1632 int proc_dointvec_minmax(struct ctl_table *table, int write,
1633 		    void *buffer, size_t *lenp, loff_t *ppos)
1634 {
1635 	return -ENOSYS;
1636 }
1637 
1638 int proc_douintvec_minmax(struct ctl_table *table, int write,
1639 			  void *buffer, size_t *lenp, loff_t *ppos)
1640 {
1641 	return -ENOSYS;
1642 }
1643 
1644 int proc_dou8vec_minmax(struct ctl_table *table, int write,
1645 			void *buffer, size_t *lenp, loff_t *ppos)
1646 {
1647 	return -ENOSYS;
1648 }
1649 
1650 int proc_dointvec_jiffies(struct ctl_table *table, int write,
1651 		    void *buffer, size_t *lenp, loff_t *ppos)
1652 {
1653 	return -ENOSYS;
1654 }
1655 
1656 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1657 		    void *buffer, size_t *lenp, loff_t *ppos)
1658 {
1659 	return -ENOSYS;
1660 }
1661 
1662 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
1663 			     void *buffer, size_t *lenp, loff_t *ppos)
1664 {
1665 	return -ENOSYS;
1666 }
1667 
1668 int proc_doulongvec_minmax(struct ctl_table *table, int write,
1669 		    void *buffer, size_t *lenp, loff_t *ppos)
1670 {
1671 	return -ENOSYS;
1672 }
1673 
1674 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1675 				      void *buffer, size_t *lenp, loff_t *ppos)
1676 {
1677 	return -ENOSYS;
1678 }
1679 
1680 int proc_do_large_bitmap(struct ctl_table *table, int write,
1681 			 void *buffer, size_t *lenp, loff_t *ppos)
1682 {
1683 	return -ENOSYS;
1684 }
1685 
1686 #endif /* CONFIG_PROC_SYSCTL */
1687 
1688 #if defined(CONFIG_SYSCTL)
1689 int proc_do_static_key(struct ctl_table *table, int write,
1690 		       void *buffer, size_t *lenp, loff_t *ppos)
1691 {
1692 	struct static_key *key = (struct static_key *)table->data;
1693 	static DEFINE_MUTEX(static_key_mutex);
1694 	int val, ret;
1695 	struct ctl_table tmp = {
1696 		.data   = &val,
1697 		.maxlen = sizeof(val),
1698 		.mode   = table->mode,
1699 		.extra1 = SYSCTL_ZERO,
1700 		.extra2 = SYSCTL_ONE,
1701 	};
1702 
1703 	if (write && !capable(CAP_SYS_ADMIN))
1704 		return -EPERM;
1705 
1706 	mutex_lock(&static_key_mutex);
1707 	val = static_key_enabled(key);
1708 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
1709 	if (write && !ret) {
1710 		if (val)
1711 			static_key_enable(key);
1712 		else
1713 			static_key_disable(key);
1714 	}
1715 	mutex_unlock(&static_key_mutex);
1716 	return ret;
1717 }
1718 
1719 static struct ctl_table kern_table[] = {
1720 	{
1721 		.procname	= "sched_child_runs_first",
1722 		.data		= &sysctl_sched_child_runs_first,
1723 		.maxlen		= sizeof(unsigned int),
1724 		.mode		= 0644,
1725 		.proc_handler	= proc_dointvec,
1726 	},
1727 #ifdef CONFIG_SCHED_DEBUG
1728 	{
1729 		.procname	= "sched_min_granularity_ns",
1730 		.data		= &sysctl_sched_min_granularity,
1731 		.maxlen		= sizeof(unsigned int),
1732 		.mode		= 0644,
1733 		.proc_handler	= sched_proc_update_handler,
1734 		.extra1		= &min_sched_granularity_ns,
1735 		.extra2		= &max_sched_granularity_ns,
1736 	},
1737 	{
1738 		.procname	= "sched_latency_ns",
1739 		.data		= &sysctl_sched_latency,
1740 		.maxlen		= sizeof(unsigned int),
1741 		.mode		= 0644,
1742 		.proc_handler	= sched_proc_update_handler,
1743 		.extra1		= &min_sched_granularity_ns,
1744 		.extra2		= &max_sched_granularity_ns,
1745 	},
1746 	{
1747 		.procname	= "sched_wakeup_granularity_ns",
1748 		.data		= &sysctl_sched_wakeup_granularity,
1749 		.maxlen		= sizeof(unsigned int),
1750 		.mode		= 0644,
1751 		.proc_handler	= sched_proc_update_handler,
1752 		.extra1		= &min_wakeup_granularity_ns,
1753 		.extra2		= &max_wakeup_granularity_ns,
1754 	},
1755 #ifdef CONFIG_SMP
1756 	{
1757 		.procname	= "sched_tunable_scaling",
1758 		.data		= &sysctl_sched_tunable_scaling,
1759 		.maxlen		= sizeof(enum sched_tunable_scaling),
1760 		.mode		= 0644,
1761 		.proc_handler	= sched_proc_update_handler,
1762 		.extra1		= &min_sched_tunable_scaling,
1763 		.extra2		= &max_sched_tunable_scaling,
1764 	},
1765 	{
1766 		.procname	= "sched_migration_cost_ns",
1767 		.data		= &sysctl_sched_migration_cost,
1768 		.maxlen		= sizeof(unsigned int),
1769 		.mode		= 0644,
1770 		.proc_handler	= proc_dointvec,
1771 	},
1772 	{
1773 		.procname	= "sched_nr_migrate",
1774 		.data		= &sysctl_sched_nr_migrate,
1775 		.maxlen		= sizeof(unsigned int),
1776 		.mode		= 0644,
1777 		.proc_handler	= proc_dointvec,
1778 	},
1779 #ifdef CONFIG_SCHEDSTATS
1780 	{
1781 		.procname	= "sched_schedstats",
1782 		.data		= NULL,
1783 		.maxlen		= sizeof(unsigned int),
1784 		.mode		= 0644,
1785 		.proc_handler	= sysctl_schedstats,
1786 		.extra1		= SYSCTL_ZERO,
1787 		.extra2		= SYSCTL_ONE,
1788 	},
1789 #endif /* CONFIG_SCHEDSTATS */
1790 #endif /* CONFIG_SMP */
1791 #ifdef CONFIG_NUMA_BALANCING
1792 	{
1793 		.procname	= "numa_balancing_scan_delay_ms",
1794 		.data		= &sysctl_numa_balancing_scan_delay,
1795 		.maxlen		= sizeof(unsigned int),
1796 		.mode		= 0644,
1797 		.proc_handler	= proc_dointvec,
1798 	},
1799 	{
1800 		.procname	= "numa_balancing_scan_period_min_ms",
1801 		.data		= &sysctl_numa_balancing_scan_period_min,
1802 		.maxlen		= sizeof(unsigned int),
1803 		.mode		= 0644,
1804 		.proc_handler	= proc_dointvec,
1805 	},
1806 	{
1807 		.procname	= "numa_balancing_scan_period_max_ms",
1808 		.data		= &sysctl_numa_balancing_scan_period_max,
1809 		.maxlen		= sizeof(unsigned int),
1810 		.mode		= 0644,
1811 		.proc_handler	= proc_dointvec,
1812 	},
1813 	{
1814 		.procname	= "numa_balancing_scan_size_mb",
1815 		.data		= &sysctl_numa_balancing_scan_size,
1816 		.maxlen		= sizeof(unsigned int),
1817 		.mode		= 0644,
1818 		.proc_handler	= proc_dointvec_minmax,
1819 		.extra1		= SYSCTL_ONE,
1820 	},
1821 	{
1822 		.procname	= "numa_balancing",
1823 		.data		= NULL, /* filled in by handler */
1824 		.maxlen		= sizeof(unsigned int),
1825 		.mode		= 0644,
1826 		.proc_handler	= sysctl_numa_balancing,
1827 		.extra1		= SYSCTL_ZERO,
1828 		.extra2		= SYSCTL_ONE,
1829 	},
1830 #endif /* CONFIG_NUMA_BALANCING */
1831 #endif /* CONFIG_SCHED_DEBUG */
1832 	{
1833 		.procname	= "sched_rt_period_us",
1834 		.data		= &sysctl_sched_rt_period,
1835 		.maxlen		= sizeof(unsigned int),
1836 		.mode		= 0644,
1837 		.proc_handler	= sched_rt_handler,
1838 	},
1839 	{
1840 		.procname	= "sched_rt_runtime_us",
1841 		.data		= &sysctl_sched_rt_runtime,
1842 		.maxlen		= sizeof(int),
1843 		.mode		= 0644,
1844 		.proc_handler	= sched_rt_handler,
1845 	},
1846 	{
1847 		.procname	= "sched_deadline_period_max_us",
1848 		.data		= &sysctl_sched_dl_period_max,
1849 		.maxlen		= sizeof(unsigned int),
1850 		.mode		= 0644,
1851 		.proc_handler	= proc_dointvec,
1852 	},
1853 	{
1854 		.procname	= "sched_deadline_period_min_us",
1855 		.data		= &sysctl_sched_dl_period_min,
1856 		.maxlen		= sizeof(unsigned int),
1857 		.mode		= 0644,
1858 		.proc_handler	= proc_dointvec,
1859 	},
1860 	{
1861 		.procname	= "sched_rr_timeslice_ms",
1862 		.data		= &sysctl_sched_rr_timeslice,
1863 		.maxlen		= sizeof(int),
1864 		.mode		= 0644,
1865 		.proc_handler	= sched_rr_handler,
1866 	},
1867 #ifdef CONFIG_UCLAMP_TASK
1868 	{
1869 		.procname	= "sched_util_clamp_min",
1870 		.data		= &sysctl_sched_uclamp_util_min,
1871 		.maxlen		= sizeof(unsigned int),
1872 		.mode		= 0644,
1873 		.proc_handler	= sysctl_sched_uclamp_handler,
1874 	},
1875 	{
1876 		.procname	= "sched_util_clamp_max",
1877 		.data		= &sysctl_sched_uclamp_util_max,
1878 		.maxlen		= sizeof(unsigned int),
1879 		.mode		= 0644,
1880 		.proc_handler	= sysctl_sched_uclamp_handler,
1881 	},
1882 	{
1883 		.procname	= "sched_util_clamp_min_rt_default",
1884 		.data		= &sysctl_sched_uclamp_util_min_rt_default,
1885 		.maxlen		= sizeof(unsigned int),
1886 		.mode		= 0644,
1887 		.proc_handler	= sysctl_sched_uclamp_handler,
1888 	},
1889 #endif
1890 #ifdef CONFIG_SCHED_AUTOGROUP
1891 	{
1892 		.procname	= "sched_autogroup_enabled",
1893 		.data		= &sysctl_sched_autogroup_enabled,
1894 		.maxlen		= sizeof(unsigned int),
1895 		.mode		= 0644,
1896 		.proc_handler	= proc_dointvec_minmax,
1897 		.extra1		= SYSCTL_ZERO,
1898 		.extra2		= SYSCTL_ONE,
1899 	},
1900 #endif
1901 #ifdef CONFIG_CFS_BANDWIDTH
1902 	{
1903 		.procname	= "sched_cfs_bandwidth_slice_us",
1904 		.data		= &sysctl_sched_cfs_bandwidth_slice,
1905 		.maxlen		= sizeof(unsigned int),
1906 		.mode		= 0644,
1907 		.proc_handler	= proc_dointvec_minmax,
1908 		.extra1		= SYSCTL_ONE,
1909 	},
1910 #endif
1911 #if defined(CONFIG_ENERGY_MODEL) && defined(CONFIG_CPU_FREQ_GOV_SCHEDUTIL)
1912 	{
1913 		.procname	= "sched_energy_aware",
1914 		.data		= &sysctl_sched_energy_aware,
1915 		.maxlen		= sizeof(unsigned int),
1916 		.mode		= 0644,
1917 		.proc_handler	= sched_energy_aware_handler,
1918 		.extra1		= SYSCTL_ZERO,
1919 		.extra2		= SYSCTL_ONE,
1920 	},
1921 #endif
1922 #ifdef CONFIG_PROVE_LOCKING
1923 	{
1924 		.procname	= "prove_locking",
1925 		.data		= &prove_locking,
1926 		.maxlen		= sizeof(int),
1927 		.mode		= 0644,
1928 		.proc_handler	= proc_dointvec,
1929 	},
1930 #endif
1931 #ifdef CONFIG_LOCK_STAT
1932 	{
1933 		.procname	= "lock_stat",
1934 		.data		= &lock_stat,
1935 		.maxlen		= sizeof(int),
1936 		.mode		= 0644,
1937 		.proc_handler	= proc_dointvec,
1938 	},
1939 #endif
1940 	{
1941 		.procname	= "panic",
1942 		.data		= &panic_timeout,
1943 		.maxlen		= sizeof(int),
1944 		.mode		= 0644,
1945 		.proc_handler	= proc_dointvec,
1946 	},
1947 #ifdef CONFIG_COREDUMP
1948 	{
1949 		.procname	= "core_uses_pid",
1950 		.data		= &core_uses_pid,
1951 		.maxlen		= sizeof(int),
1952 		.mode		= 0644,
1953 		.proc_handler	= proc_dointvec,
1954 	},
1955 	{
1956 		.procname	= "core_pattern",
1957 		.data		= core_pattern,
1958 		.maxlen		= CORENAME_MAX_SIZE,
1959 		.mode		= 0644,
1960 		.proc_handler	= proc_dostring_coredump,
1961 	},
1962 	{
1963 		.procname	= "core_pipe_limit",
1964 		.data		= &core_pipe_limit,
1965 		.maxlen		= sizeof(unsigned int),
1966 		.mode		= 0644,
1967 		.proc_handler	= proc_dointvec,
1968 	},
1969 #endif
1970 #ifdef CONFIG_PROC_SYSCTL
1971 	{
1972 		.procname	= "tainted",
1973 		.maxlen 	= sizeof(long),
1974 		.mode		= 0644,
1975 		.proc_handler	= proc_taint,
1976 	},
1977 	{
1978 		.procname	= "sysctl_writes_strict",
1979 		.data		= &sysctl_writes_strict,
1980 		.maxlen		= sizeof(int),
1981 		.mode		= 0644,
1982 		.proc_handler	= proc_dointvec_minmax,
1983 		.extra1		= &neg_one,
1984 		.extra2		= SYSCTL_ONE,
1985 	},
1986 #endif
1987 #ifdef CONFIG_LATENCYTOP
1988 	{
1989 		.procname	= "latencytop",
1990 		.data		= &latencytop_enabled,
1991 		.maxlen		= sizeof(int),
1992 		.mode		= 0644,
1993 		.proc_handler	= sysctl_latencytop,
1994 	},
1995 #endif
1996 #ifdef CONFIG_BLK_DEV_INITRD
1997 	{
1998 		.procname	= "real-root-dev",
1999 		.data		= &real_root_dev,
2000 		.maxlen		= sizeof(int),
2001 		.mode		= 0644,
2002 		.proc_handler	= proc_dointvec,
2003 	},
2004 #endif
2005 	{
2006 		.procname	= "print-fatal-signals",
2007 		.data		= &print_fatal_signals,
2008 		.maxlen		= sizeof(int),
2009 		.mode		= 0644,
2010 		.proc_handler	= proc_dointvec,
2011 	},
2012 #ifdef CONFIG_SPARC
2013 	{
2014 		.procname	= "reboot-cmd",
2015 		.data		= reboot_command,
2016 		.maxlen		= 256,
2017 		.mode		= 0644,
2018 		.proc_handler	= proc_dostring,
2019 	},
2020 	{
2021 		.procname	= "stop-a",
2022 		.data		= &stop_a_enabled,
2023 		.maxlen		= sizeof (int),
2024 		.mode		= 0644,
2025 		.proc_handler	= proc_dointvec,
2026 	},
2027 	{
2028 		.procname	= "scons-poweroff",
2029 		.data		= &scons_pwroff,
2030 		.maxlen		= sizeof (int),
2031 		.mode		= 0644,
2032 		.proc_handler	= proc_dointvec,
2033 	},
2034 #endif
2035 #ifdef CONFIG_SPARC64
2036 	{
2037 		.procname	= "tsb-ratio",
2038 		.data		= &sysctl_tsb_ratio,
2039 		.maxlen		= sizeof (int),
2040 		.mode		= 0644,
2041 		.proc_handler	= proc_dointvec,
2042 	},
2043 #endif
2044 #ifdef CONFIG_PARISC
2045 	{
2046 		.procname	= "soft-power",
2047 		.data		= &pwrsw_enabled,
2048 		.maxlen		= sizeof (int),
2049 		.mode		= 0644,
2050 		.proc_handler	= proc_dointvec,
2051 	},
2052 #endif
2053 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
2054 	{
2055 		.procname	= "unaligned-trap",
2056 		.data		= &unaligned_enabled,
2057 		.maxlen		= sizeof (int),
2058 		.mode		= 0644,
2059 		.proc_handler	= proc_dointvec,
2060 	},
2061 #endif
2062 	{
2063 		.procname	= "ctrl-alt-del",
2064 		.data		= &C_A_D,
2065 		.maxlen		= sizeof(int),
2066 		.mode		= 0644,
2067 		.proc_handler	= proc_dointvec,
2068 	},
2069 #ifdef CONFIG_FUNCTION_TRACER
2070 	{
2071 		.procname	= "ftrace_enabled",
2072 		.data		= &ftrace_enabled,
2073 		.maxlen		= sizeof(int),
2074 		.mode		= 0644,
2075 		.proc_handler	= ftrace_enable_sysctl,
2076 	},
2077 #endif
2078 #ifdef CONFIG_STACK_TRACER
2079 	{
2080 		.procname	= "stack_tracer_enabled",
2081 		.data		= &stack_tracer_enabled,
2082 		.maxlen		= sizeof(int),
2083 		.mode		= 0644,
2084 		.proc_handler	= stack_trace_sysctl,
2085 	},
2086 #endif
2087 #ifdef CONFIG_TRACING
2088 	{
2089 		.procname	= "ftrace_dump_on_oops",
2090 		.data		= &ftrace_dump_on_oops,
2091 		.maxlen		= sizeof(int),
2092 		.mode		= 0644,
2093 		.proc_handler	= proc_dointvec,
2094 	},
2095 	{
2096 		.procname	= "traceoff_on_warning",
2097 		.data		= &__disable_trace_on_warning,
2098 		.maxlen		= sizeof(__disable_trace_on_warning),
2099 		.mode		= 0644,
2100 		.proc_handler	= proc_dointvec,
2101 	},
2102 	{
2103 		.procname	= "tracepoint_printk",
2104 		.data		= &tracepoint_printk,
2105 		.maxlen		= sizeof(tracepoint_printk),
2106 		.mode		= 0644,
2107 		.proc_handler	= tracepoint_printk_sysctl,
2108 	},
2109 #endif
2110 #ifdef CONFIG_KEXEC_CORE
2111 	{
2112 		.procname	= "kexec_load_disabled",
2113 		.data		= &kexec_load_disabled,
2114 		.maxlen		= sizeof(int),
2115 		.mode		= 0644,
2116 		/* only handle a transition from default "0" to "1" */
2117 		.proc_handler	= proc_dointvec_minmax,
2118 		.extra1		= SYSCTL_ONE,
2119 		.extra2		= SYSCTL_ONE,
2120 	},
2121 #endif
2122 #ifdef CONFIG_MODULES
2123 	{
2124 		.procname	= "modprobe",
2125 		.data		= &modprobe_path,
2126 		.maxlen		= KMOD_PATH_LEN,
2127 		.mode		= 0644,
2128 		.proc_handler	= proc_dostring,
2129 	},
2130 	{
2131 		.procname	= "modules_disabled",
2132 		.data		= &modules_disabled,
2133 		.maxlen		= sizeof(int),
2134 		.mode		= 0644,
2135 		/* only handle a transition from default "0" to "1" */
2136 		.proc_handler	= proc_dointvec_minmax,
2137 		.extra1		= SYSCTL_ONE,
2138 		.extra2		= SYSCTL_ONE,
2139 	},
2140 #endif
2141 #ifdef CONFIG_UEVENT_HELPER
2142 	{
2143 		.procname	= "hotplug",
2144 		.data		= &uevent_helper,
2145 		.maxlen		= UEVENT_HELPER_PATH_LEN,
2146 		.mode		= 0644,
2147 		.proc_handler	= proc_dostring,
2148 	},
2149 #endif
2150 #ifdef CONFIG_CHR_DEV_SG
2151 	{
2152 		.procname	= "sg-big-buff",
2153 		.data		= &sg_big_buff,
2154 		.maxlen		= sizeof (int),
2155 		.mode		= 0444,
2156 		.proc_handler	= proc_dointvec,
2157 	},
2158 #endif
2159 #ifdef CONFIG_BSD_PROCESS_ACCT
2160 	{
2161 		.procname	= "acct",
2162 		.data		= &acct_parm,
2163 		.maxlen		= 3*sizeof(int),
2164 		.mode		= 0644,
2165 		.proc_handler	= proc_dointvec,
2166 	},
2167 #endif
2168 #ifdef CONFIG_MAGIC_SYSRQ
2169 	{
2170 		.procname	= "sysrq",
2171 		.data		= NULL,
2172 		.maxlen		= sizeof (int),
2173 		.mode		= 0644,
2174 		.proc_handler	= sysrq_sysctl_handler,
2175 	},
2176 #endif
2177 #ifdef CONFIG_PROC_SYSCTL
2178 	{
2179 		.procname	= "cad_pid",
2180 		.data		= NULL,
2181 		.maxlen		= sizeof (int),
2182 		.mode		= 0600,
2183 		.proc_handler	= proc_do_cad_pid,
2184 	},
2185 #endif
2186 	{
2187 		.procname	= "threads-max",
2188 		.data		= NULL,
2189 		.maxlen		= sizeof(int),
2190 		.mode		= 0644,
2191 		.proc_handler	= sysctl_max_threads,
2192 	},
2193 	{
2194 		.procname	= "random",
2195 		.mode		= 0555,
2196 		.child		= random_table,
2197 	},
2198 	{
2199 		.procname	= "usermodehelper",
2200 		.mode		= 0555,
2201 		.child		= usermodehelper_table,
2202 	},
2203 #ifdef CONFIG_FW_LOADER_USER_HELPER
2204 	{
2205 		.procname	= "firmware_config",
2206 		.mode		= 0555,
2207 		.child		= firmware_config_table,
2208 	},
2209 #endif
2210 	{
2211 		.procname	= "overflowuid",
2212 		.data		= &overflowuid,
2213 		.maxlen		= sizeof(int),
2214 		.mode		= 0644,
2215 		.proc_handler	= proc_dointvec_minmax,
2216 		.extra1		= &minolduid,
2217 		.extra2		= &maxolduid,
2218 	},
2219 	{
2220 		.procname	= "overflowgid",
2221 		.data		= &overflowgid,
2222 		.maxlen		= sizeof(int),
2223 		.mode		= 0644,
2224 		.proc_handler	= proc_dointvec_minmax,
2225 		.extra1		= &minolduid,
2226 		.extra2		= &maxolduid,
2227 	},
2228 #ifdef CONFIG_S390
2229 	{
2230 		.procname	= "userprocess_debug",
2231 		.data		= &show_unhandled_signals,
2232 		.maxlen		= sizeof(int),
2233 		.mode		= 0644,
2234 		.proc_handler	= proc_dointvec,
2235 	},
2236 #endif
2237 #ifdef CONFIG_SMP
2238 	{
2239 		.procname	= "oops_all_cpu_backtrace",
2240 		.data		= &sysctl_oops_all_cpu_backtrace,
2241 		.maxlen		= sizeof(int),
2242 		.mode		= 0644,
2243 		.proc_handler	= proc_dointvec_minmax,
2244 		.extra1		= SYSCTL_ZERO,
2245 		.extra2		= SYSCTL_ONE,
2246 	},
2247 #endif /* CONFIG_SMP */
2248 	{
2249 		.procname	= "pid_max",
2250 		.data		= &pid_max,
2251 		.maxlen		= sizeof (int),
2252 		.mode		= 0644,
2253 		.proc_handler	= proc_dointvec_minmax,
2254 		.extra1		= &pid_max_min,
2255 		.extra2		= &pid_max_max,
2256 	},
2257 	{
2258 		.procname	= "panic_on_oops",
2259 		.data		= &panic_on_oops,
2260 		.maxlen		= sizeof(int),
2261 		.mode		= 0644,
2262 		.proc_handler	= proc_dointvec,
2263 	},
2264 	{
2265 		.procname	= "panic_print",
2266 		.data		= &panic_print,
2267 		.maxlen		= sizeof(unsigned long),
2268 		.mode		= 0644,
2269 		.proc_handler	= proc_doulongvec_minmax,
2270 	},
2271 #if defined CONFIG_PRINTK
2272 	{
2273 		.procname	= "printk",
2274 		.data		= &console_loglevel,
2275 		.maxlen		= 4*sizeof(int),
2276 		.mode		= 0644,
2277 		.proc_handler	= proc_dointvec,
2278 	},
2279 	{
2280 		.procname	= "printk_ratelimit",
2281 		.data		= &printk_ratelimit_state.interval,
2282 		.maxlen		= sizeof(int),
2283 		.mode		= 0644,
2284 		.proc_handler	= proc_dointvec_jiffies,
2285 	},
2286 	{
2287 		.procname	= "printk_ratelimit_burst",
2288 		.data		= &printk_ratelimit_state.burst,
2289 		.maxlen		= sizeof(int),
2290 		.mode		= 0644,
2291 		.proc_handler	= proc_dointvec,
2292 	},
2293 	{
2294 		.procname	= "printk_delay",
2295 		.data		= &printk_delay_msec,
2296 		.maxlen		= sizeof(int),
2297 		.mode		= 0644,
2298 		.proc_handler	= proc_dointvec_minmax,
2299 		.extra1		= SYSCTL_ZERO,
2300 		.extra2		= &ten_thousand,
2301 	},
2302 	{
2303 		.procname	= "printk_devkmsg",
2304 		.data		= devkmsg_log_str,
2305 		.maxlen		= DEVKMSG_STR_MAX_SIZE,
2306 		.mode		= 0644,
2307 		.proc_handler	= devkmsg_sysctl_set_loglvl,
2308 	},
2309 	{
2310 		.procname	= "dmesg_restrict",
2311 		.data		= &dmesg_restrict,
2312 		.maxlen		= sizeof(int),
2313 		.mode		= 0644,
2314 		.proc_handler	= proc_dointvec_minmax_sysadmin,
2315 		.extra1		= SYSCTL_ZERO,
2316 		.extra2		= SYSCTL_ONE,
2317 	},
2318 	{
2319 		.procname	= "kptr_restrict",
2320 		.data		= &kptr_restrict,
2321 		.maxlen		= sizeof(int),
2322 		.mode		= 0644,
2323 		.proc_handler	= proc_dointvec_minmax_sysadmin,
2324 		.extra1		= SYSCTL_ZERO,
2325 		.extra2		= &two,
2326 	},
2327 #endif
2328 	{
2329 		.procname	= "ngroups_max",
2330 		.data		= &ngroups_max,
2331 		.maxlen		= sizeof (int),
2332 		.mode		= 0444,
2333 		.proc_handler	= proc_dointvec,
2334 	},
2335 	{
2336 		.procname	= "cap_last_cap",
2337 		.data		= (void *)&cap_last_cap,
2338 		.maxlen		= sizeof(int),
2339 		.mode		= 0444,
2340 		.proc_handler	= proc_dointvec,
2341 	},
2342 #if defined(CONFIG_LOCKUP_DETECTOR)
2343 	{
2344 		.procname       = "watchdog",
2345 		.data		= &watchdog_user_enabled,
2346 		.maxlen		= sizeof(int),
2347 		.mode		= 0644,
2348 		.proc_handler   = proc_watchdog,
2349 		.extra1		= SYSCTL_ZERO,
2350 		.extra2		= SYSCTL_ONE,
2351 	},
2352 	{
2353 		.procname	= "watchdog_thresh",
2354 		.data		= &watchdog_thresh,
2355 		.maxlen		= sizeof(int),
2356 		.mode		= 0644,
2357 		.proc_handler	= proc_watchdog_thresh,
2358 		.extra1		= SYSCTL_ZERO,
2359 		.extra2		= &sixty,
2360 	},
2361 	{
2362 		.procname       = "nmi_watchdog",
2363 		.data		= &nmi_watchdog_user_enabled,
2364 		.maxlen		= sizeof(int),
2365 		.mode		= NMI_WATCHDOG_SYSCTL_PERM,
2366 		.proc_handler   = proc_nmi_watchdog,
2367 		.extra1		= SYSCTL_ZERO,
2368 		.extra2		= SYSCTL_ONE,
2369 	},
2370 	{
2371 		.procname	= "watchdog_cpumask",
2372 		.data		= &watchdog_cpumask_bits,
2373 		.maxlen		= NR_CPUS,
2374 		.mode		= 0644,
2375 		.proc_handler	= proc_watchdog_cpumask,
2376 	},
2377 #ifdef CONFIG_SOFTLOCKUP_DETECTOR
2378 	{
2379 		.procname       = "soft_watchdog",
2380 		.data		= &soft_watchdog_user_enabled,
2381 		.maxlen		= sizeof(int),
2382 		.mode		= 0644,
2383 		.proc_handler   = proc_soft_watchdog,
2384 		.extra1		= SYSCTL_ZERO,
2385 		.extra2		= SYSCTL_ONE,
2386 	},
2387 	{
2388 		.procname	= "softlockup_panic",
2389 		.data		= &softlockup_panic,
2390 		.maxlen		= sizeof(int),
2391 		.mode		= 0644,
2392 		.proc_handler	= proc_dointvec_minmax,
2393 		.extra1		= SYSCTL_ZERO,
2394 		.extra2		= SYSCTL_ONE,
2395 	},
2396 #ifdef CONFIG_SMP
2397 	{
2398 		.procname	= "softlockup_all_cpu_backtrace",
2399 		.data		= &sysctl_softlockup_all_cpu_backtrace,
2400 		.maxlen		= sizeof(int),
2401 		.mode		= 0644,
2402 		.proc_handler	= proc_dointvec_minmax,
2403 		.extra1		= SYSCTL_ZERO,
2404 		.extra2		= SYSCTL_ONE,
2405 	},
2406 #endif /* CONFIG_SMP */
2407 #endif
2408 #ifdef CONFIG_HARDLOCKUP_DETECTOR
2409 	{
2410 		.procname	= "hardlockup_panic",
2411 		.data		= &hardlockup_panic,
2412 		.maxlen		= sizeof(int),
2413 		.mode		= 0644,
2414 		.proc_handler	= proc_dointvec_minmax,
2415 		.extra1		= SYSCTL_ZERO,
2416 		.extra2		= SYSCTL_ONE,
2417 	},
2418 #ifdef CONFIG_SMP
2419 	{
2420 		.procname	= "hardlockup_all_cpu_backtrace",
2421 		.data		= &sysctl_hardlockup_all_cpu_backtrace,
2422 		.maxlen		= sizeof(int),
2423 		.mode		= 0644,
2424 		.proc_handler	= proc_dointvec_minmax,
2425 		.extra1		= SYSCTL_ZERO,
2426 		.extra2		= SYSCTL_ONE,
2427 	},
2428 #endif /* CONFIG_SMP */
2429 #endif
2430 #endif
2431 
2432 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
2433 	{
2434 		.procname       = "unknown_nmi_panic",
2435 		.data           = &unknown_nmi_panic,
2436 		.maxlen         = sizeof (int),
2437 		.mode           = 0644,
2438 		.proc_handler   = proc_dointvec,
2439 	},
2440 #endif
2441 
2442 #if (defined(CONFIG_X86_32) || defined(CONFIG_PARISC)) && \
2443 	defined(CONFIG_DEBUG_STACKOVERFLOW)
2444 	{
2445 		.procname	= "panic_on_stackoverflow",
2446 		.data		= &sysctl_panic_on_stackoverflow,
2447 		.maxlen		= sizeof(int),
2448 		.mode		= 0644,
2449 		.proc_handler	= proc_dointvec,
2450 	},
2451 #endif
2452 #if defined(CONFIG_X86)
2453 	{
2454 		.procname	= "panic_on_unrecovered_nmi",
2455 		.data		= &panic_on_unrecovered_nmi,
2456 		.maxlen		= sizeof(int),
2457 		.mode		= 0644,
2458 		.proc_handler	= proc_dointvec,
2459 	},
2460 	{
2461 		.procname	= "panic_on_io_nmi",
2462 		.data		= &panic_on_io_nmi,
2463 		.maxlen		= sizeof(int),
2464 		.mode		= 0644,
2465 		.proc_handler	= proc_dointvec,
2466 	},
2467 	{
2468 		.procname	= "bootloader_type",
2469 		.data		= &bootloader_type,
2470 		.maxlen		= sizeof (int),
2471 		.mode		= 0444,
2472 		.proc_handler	= proc_dointvec,
2473 	},
2474 	{
2475 		.procname	= "bootloader_version",
2476 		.data		= &bootloader_version,
2477 		.maxlen		= sizeof (int),
2478 		.mode		= 0444,
2479 		.proc_handler	= proc_dointvec,
2480 	},
2481 	{
2482 		.procname	= "io_delay_type",
2483 		.data		= &io_delay_type,
2484 		.maxlen		= sizeof(int),
2485 		.mode		= 0644,
2486 		.proc_handler	= proc_dointvec,
2487 	},
2488 #endif
2489 #if defined(CONFIG_MMU)
2490 	{
2491 		.procname	= "randomize_va_space",
2492 		.data		= &randomize_va_space,
2493 		.maxlen		= sizeof(int),
2494 		.mode		= 0644,
2495 		.proc_handler	= proc_dointvec,
2496 	},
2497 #endif
2498 #if defined(CONFIG_S390) && defined(CONFIG_SMP)
2499 	{
2500 		.procname	= "spin_retry",
2501 		.data		= &spin_retry,
2502 		.maxlen		= sizeof (int),
2503 		.mode		= 0644,
2504 		.proc_handler	= proc_dointvec,
2505 	},
2506 #endif
2507 #if	defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
2508 	{
2509 		.procname	= "acpi_video_flags",
2510 		.data		= &acpi_realmode_flags,
2511 		.maxlen		= sizeof (unsigned long),
2512 		.mode		= 0644,
2513 		.proc_handler	= proc_doulongvec_minmax,
2514 	},
2515 #endif
2516 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
2517 	{
2518 		.procname	= "ignore-unaligned-usertrap",
2519 		.data		= &no_unaligned_warning,
2520 		.maxlen		= sizeof (int),
2521 		.mode		= 0644,
2522 		.proc_handler	= proc_dointvec,
2523 	},
2524 #endif
2525 #ifdef CONFIG_IA64
2526 	{
2527 		.procname	= "unaligned-dump-stack",
2528 		.data		= &unaligned_dump_stack,
2529 		.maxlen		= sizeof (int),
2530 		.mode		= 0644,
2531 		.proc_handler	= proc_dointvec,
2532 	},
2533 #endif
2534 #ifdef CONFIG_DETECT_HUNG_TASK
2535 #ifdef CONFIG_SMP
2536 	{
2537 		.procname	= "hung_task_all_cpu_backtrace",
2538 		.data		= &sysctl_hung_task_all_cpu_backtrace,
2539 		.maxlen		= sizeof(int),
2540 		.mode		= 0644,
2541 		.proc_handler	= proc_dointvec_minmax,
2542 		.extra1		= SYSCTL_ZERO,
2543 		.extra2		= SYSCTL_ONE,
2544 	},
2545 #endif /* CONFIG_SMP */
2546 	{
2547 		.procname	= "hung_task_panic",
2548 		.data		= &sysctl_hung_task_panic,
2549 		.maxlen		= sizeof(int),
2550 		.mode		= 0644,
2551 		.proc_handler	= proc_dointvec_minmax,
2552 		.extra1		= SYSCTL_ZERO,
2553 		.extra2		= SYSCTL_ONE,
2554 	},
2555 	{
2556 		.procname	= "hung_task_check_count",
2557 		.data		= &sysctl_hung_task_check_count,
2558 		.maxlen		= sizeof(int),
2559 		.mode		= 0644,
2560 		.proc_handler	= proc_dointvec_minmax,
2561 		.extra1		= SYSCTL_ZERO,
2562 	},
2563 	{
2564 		.procname	= "hung_task_timeout_secs",
2565 		.data		= &sysctl_hung_task_timeout_secs,
2566 		.maxlen		= sizeof(unsigned long),
2567 		.mode		= 0644,
2568 		.proc_handler	= proc_dohung_task_timeout_secs,
2569 		.extra2		= &hung_task_timeout_max,
2570 	},
2571 	{
2572 		.procname	= "hung_task_check_interval_secs",
2573 		.data		= &sysctl_hung_task_check_interval_secs,
2574 		.maxlen		= sizeof(unsigned long),
2575 		.mode		= 0644,
2576 		.proc_handler	= proc_dohung_task_timeout_secs,
2577 		.extra2		= &hung_task_timeout_max,
2578 	},
2579 	{
2580 		.procname	= "hung_task_warnings",
2581 		.data		= &sysctl_hung_task_warnings,
2582 		.maxlen		= sizeof(int),
2583 		.mode		= 0644,
2584 		.proc_handler	= proc_dointvec_minmax,
2585 		.extra1		= &neg_one,
2586 	},
2587 #endif
2588 #ifdef CONFIG_RT_MUTEXES
2589 	{
2590 		.procname	= "max_lock_depth",
2591 		.data		= &max_lock_depth,
2592 		.maxlen		= sizeof(int),
2593 		.mode		= 0644,
2594 		.proc_handler	= proc_dointvec,
2595 	},
2596 #endif
2597 	{
2598 		.procname	= "poweroff_cmd",
2599 		.data		= &poweroff_cmd,
2600 		.maxlen		= POWEROFF_CMD_PATH_LEN,
2601 		.mode		= 0644,
2602 		.proc_handler	= proc_dostring,
2603 	},
2604 #ifdef CONFIG_KEYS
2605 	{
2606 		.procname	= "keys",
2607 		.mode		= 0555,
2608 		.child		= key_sysctls,
2609 	},
2610 #endif
2611 #ifdef CONFIG_PERF_EVENTS
2612 	/*
2613 	 * User-space scripts rely on the existence of this file
2614 	 * as a feature check for perf_events being enabled.
2615 	 *
2616 	 * So it's an ABI, do not remove!
2617 	 */
2618 	{
2619 		.procname	= "perf_event_paranoid",
2620 		.data		= &sysctl_perf_event_paranoid,
2621 		.maxlen		= sizeof(sysctl_perf_event_paranoid),
2622 		.mode		= 0644,
2623 		.proc_handler	= proc_dointvec,
2624 	},
2625 	{
2626 		.procname	= "perf_event_mlock_kb",
2627 		.data		= &sysctl_perf_event_mlock,
2628 		.maxlen		= sizeof(sysctl_perf_event_mlock),
2629 		.mode		= 0644,
2630 		.proc_handler	= proc_dointvec,
2631 	},
2632 	{
2633 		.procname	= "perf_event_max_sample_rate",
2634 		.data		= &sysctl_perf_event_sample_rate,
2635 		.maxlen		= sizeof(sysctl_perf_event_sample_rate),
2636 		.mode		= 0644,
2637 		.proc_handler	= perf_proc_update_handler,
2638 		.extra1		= SYSCTL_ONE,
2639 	},
2640 	{
2641 		.procname	= "perf_cpu_time_max_percent",
2642 		.data		= &sysctl_perf_cpu_time_max_percent,
2643 		.maxlen		= sizeof(sysctl_perf_cpu_time_max_percent),
2644 		.mode		= 0644,
2645 		.proc_handler	= perf_cpu_time_max_percent_handler,
2646 		.extra1		= SYSCTL_ZERO,
2647 		.extra2		= &one_hundred,
2648 	},
2649 	{
2650 		.procname	= "perf_event_max_stack",
2651 		.data		= &sysctl_perf_event_max_stack,
2652 		.maxlen		= sizeof(sysctl_perf_event_max_stack),
2653 		.mode		= 0644,
2654 		.proc_handler	= perf_event_max_stack_handler,
2655 		.extra1		= SYSCTL_ZERO,
2656 		.extra2		= &six_hundred_forty_kb,
2657 	},
2658 	{
2659 		.procname	= "perf_event_max_contexts_per_stack",
2660 		.data		= &sysctl_perf_event_max_contexts_per_stack,
2661 		.maxlen		= sizeof(sysctl_perf_event_max_contexts_per_stack),
2662 		.mode		= 0644,
2663 		.proc_handler	= perf_event_max_stack_handler,
2664 		.extra1		= SYSCTL_ZERO,
2665 		.extra2		= &one_thousand,
2666 	},
2667 #endif
2668 	{
2669 		.procname	= "panic_on_warn",
2670 		.data		= &panic_on_warn,
2671 		.maxlen		= sizeof(int),
2672 		.mode		= 0644,
2673 		.proc_handler	= proc_dointvec_minmax,
2674 		.extra1		= SYSCTL_ZERO,
2675 		.extra2		= SYSCTL_ONE,
2676 	},
2677 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
2678 	{
2679 		.procname	= "timer_migration",
2680 		.data		= &sysctl_timer_migration,
2681 		.maxlen		= sizeof(unsigned int),
2682 		.mode		= 0644,
2683 		.proc_handler	= timer_migration_handler,
2684 		.extra1		= SYSCTL_ZERO,
2685 		.extra2		= SYSCTL_ONE,
2686 	},
2687 #endif
2688 #ifdef CONFIG_BPF_SYSCALL
2689 	{
2690 		.procname	= "unprivileged_bpf_disabled",
2691 		.data		= &sysctl_unprivileged_bpf_disabled,
2692 		.maxlen		= sizeof(sysctl_unprivileged_bpf_disabled),
2693 		.mode		= 0644,
2694 		/* only handle a transition from default "0" to "1" */
2695 		.proc_handler	= proc_dointvec_minmax,
2696 		.extra1		= SYSCTL_ONE,
2697 		.extra2		= SYSCTL_ONE,
2698 	},
2699 	{
2700 		.procname	= "bpf_stats_enabled",
2701 		.data		= &bpf_stats_enabled_key.key,
2702 		.maxlen		= sizeof(bpf_stats_enabled_key),
2703 		.mode		= 0644,
2704 		.proc_handler	= bpf_stats_handler,
2705 	},
2706 #endif
2707 #if defined(CONFIG_TREE_RCU)
2708 	{
2709 		.procname	= "panic_on_rcu_stall",
2710 		.data		= &sysctl_panic_on_rcu_stall,
2711 		.maxlen		= sizeof(sysctl_panic_on_rcu_stall),
2712 		.mode		= 0644,
2713 		.proc_handler	= proc_dointvec_minmax,
2714 		.extra1		= SYSCTL_ZERO,
2715 		.extra2		= SYSCTL_ONE,
2716 	},
2717 #endif
2718 #if defined(CONFIG_TREE_RCU)
2719 	{
2720 		.procname	= "max_rcu_stall_to_panic",
2721 		.data		= &sysctl_max_rcu_stall_to_panic,
2722 		.maxlen		= sizeof(sysctl_max_rcu_stall_to_panic),
2723 		.mode		= 0644,
2724 		.proc_handler	= proc_dointvec_minmax,
2725 		.extra1		= SYSCTL_ONE,
2726 		.extra2		= SYSCTL_INT_MAX,
2727 	},
2728 #endif
2729 #ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
2730 	{
2731 		.procname	= "stack_erasing",
2732 		.data		= NULL,
2733 		.maxlen		= sizeof(int),
2734 		.mode		= 0600,
2735 		.proc_handler	= stack_erasing_sysctl,
2736 		.extra1		= SYSCTL_ZERO,
2737 		.extra2		= SYSCTL_ONE,
2738 	},
2739 #endif
2740 	{ }
2741 };
2742 
2743 static struct ctl_table vm_table[] = {
2744 	{
2745 		.procname	= "overcommit_memory",
2746 		.data		= &sysctl_overcommit_memory,
2747 		.maxlen		= sizeof(sysctl_overcommit_memory),
2748 		.mode		= 0644,
2749 		.proc_handler	= overcommit_policy_handler,
2750 		.extra1		= SYSCTL_ZERO,
2751 		.extra2		= &two,
2752 	},
2753 	{
2754 		.procname	= "panic_on_oom",
2755 		.data		= &sysctl_panic_on_oom,
2756 		.maxlen		= sizeof(sysctl_panic_on_oom),
2757 		.mode		= 0644,
2758 		.proc_handler	= proc_dointvec_minmax,
2759 		.extra1		= SYSCTL_ZERO,
2760 		.extra2		= &two,
2761 	},
2762 	{
2763 		.procname	= "oom_kill_allocating_task",
2764 		.data		= &sysctl_oom_kill_allocating_task,
2765 		.maxlen		= sizeof(sysctl_oom_kill_allocating_task),
2766 		.mode		= 0644,
2767 		.proc_handler	= proc_dointvec,
2768 	},
2769 	{
2770 		.procname	= "oom_dump_tasks",
2771 		.data		= &sysctl_oom_dump_tasks,
2772 		.maxlen		= sizeof(sysctl_oom_dump_tasks),
2773 		.mode		= 0644,
2774 		.proc_handler	= proc_dointvec,
2775 	},
2776 	{
2777 		.procname	= "overcommit_ratio",
2778 		.data		= &sysctl_overcommit_ratio,
2779 		.maxlen		= sizeof(sysctl_overcommit_ratio),
2780 		.mode		= 0644,
2781 		.proc_handler	= overcommit_ratio_handler,
2782 	},
2783 	{
2784 		.procname	= "overcommit_kbytes",
2785 		.data		= &sysctl_overcommit_kbytes,
2786 		.maxlen		= sizeof(sysctl_overcommit_kbytes),
2787 		.mode		= 0644,
2788 		.proc_handler	= overcommit_kbytes_handler,
2789 	},
2790 	{
2791 		.procname	= "page-cluster",
2792 		.data		= &page_cluster,
2793 		.maxlen		= sizeof(int),
2794 		.mode		= 0644,
2795 		.proc_handler	= proc_dointvec_minmax,
2796 		.extra1		= SYSCTL_ZERO,
2797 	},
2798 	{
2799 		.procname	= "dirty_background_ratio",
2800 		.data		= &dirty_background_ratio,
2801 		.maxlen		= sizeof(dirty_background_ratio),
2802 		.mode		= 0644,
2803 		.proc_handler	= dirty_background_ratio_handler,
2804 		.extra1		= SYSCTL_ZERO,
2805 		.extra2		= &one_hundred,
2806 	},
2807 	{
2808 		.procname	= "dirty_background_bytes",
2809 		.data		= &dirty_background_bytes,
2810 		.maxlen		= sizeof(dirty_background_bytes),
2811 		.mode		= 0644,
2812 		.proc_handler	= dirty_background_bytes_handler,
2813 		.extra1		= &one_ul,
2814 	},
2815 	{
2816 		.procname	= "dirty_ratio",
2817 		.data		= &vm_dirty_ratio,
2818 		.maxlen		= sizeof(vm_dirty_ratio),
2819 		.mode		= 0644,
2820 		.proc_handler	= dirty_ratio_handler,
2821 		.extra1		= SYSCTL_ZERO,
2822 		.extra2		= &one_hundred,
2823 	},
2824 	{
2825 		.procname	= "dirty_bytes",
2826 		.data		= &vm_dirty_bytes,
2827 		.maxlen		= sizeof(vm_dirty_bytes),
2828 		.mode		= 0644,
2829 		.proc_handler	= dirty_bytes_handler,
2830 		.extra1		= &dirty_bytes_min,
2831 	},
2832 	{
2833 		.procname	= "dirty_writeback_centisecs",
2834 		.data		= &dirty_writeback_interval,
2835 		.maxlen		= sizeof(dirty_writeback_interval),
2836 		.mode		= 0644,
2837 		.proc_handler	= dirty_writeback_centisecs_handler,
2838 	},
2839 	{
2840 		.procname	= "dirty_expire_centisecs",
2841 		.data		= &dirty_expire_interval,
2842 		.maxlen		= sizeof(dirty_expire_interval),
2843 		.mode		= 0644,
2844 		.proc_handler	= proc_dointvec_minmax,
2845 		.extra1		= SYSCTL_ZERO,
2846 	},
2847 	{
2848 		.procname	= "dirtytime_expire_seconds",
2849 		.data		= &dirtytime_expire_interval,
2850 		.maxlen		= sizeof(dirtytime_expire_interval),
2851 		.mode		= 0644,
2852 		.proc_handler	= dirtytime_interval_handler,
2853 		.extra1		= SYSCTL_ZERO,
2854 	},
2855 	{
2856 		.procname	= "swappiness",
2857 		.data		= &vm_swappiness,
2858 		.maxlen		= sizeof(vm_swappiness),
2859 		.mode		= 0644,
2860 		.proc_handler	= proc_dointvec_minmax,
2861 		.extra1		= SYSCTL_ZERO,
2862 		.extra2		= &two_hundred,
2863 	},
2864 #ifdef CONFIG_HUGETLB_PAGE
2865 	{
2866 		.procname	= "nr_hugepages",
2867 		.data		= NULL,
2868 		.maxlen		= sizeof(unsigned long),
2869 		.mode		= 0644,
2870 		.proc_handler	= hugetlb_sysctl_handler,
2871 	},
2872 #ifdef CONFIG_NUMA
2873 	{
2874 		.procname       = "nr_hugepages_mempolicy",
2875 		.data           = NULL,
2876 		.maxlen         = sizeof(unsigned long),
2877 		.mode           = 0644,
2878 		.proc_handler   = &hugetlb_mempolicy_sysctl_handler,
2879 	},
2880 	{
2881 		.procname		= "numa_stat",
2882 		.data			= &sysctl_vm_numa_stat,
2883 		.maxlen			= sizeof(int),
2884 		.mode			= 0644,
2885 		.proc_handler	= sysctl_vm_numa_stat_handler,
2886 		.extra1			= SYSCTL_ZERO,
2887 		.extra2			= SYSCTL_ONE,
2888 	},
2889 #endif
2890 	 {
2891 		.procname	= "hugetlb_shm_group",
2892 		.data		= &sysctl_hugetlb_shm_group,
2893 		.maxlen		= sizeof(gid_t),
2894 		.mode		= 0644,
2895 		.proc_handler	= proc_dointvec,
2896 	 },
2897 	{
2898 		.procname	= "nr_overcommit_hugepages",
2899 		.data		= NULL,
2900 		.maxlen		= sizeof(unsigned long),
2901 		.mode		= 0644,
2902 		.proc_handler	= hugetlb_overcommit_handler,
2903 	},
2904 #endif
2905 	{
2906 		.procname	= "lowmem_reserve_ratio",
2907 		.data		= &sysctl_lowmem_reserve_ratio,
2908 		.maxlen		= sizeof(sysctl_lowmem_reserve_ratio),
2909 		.mode		= 0644,
2910 		.proc_handler	= lowmem_reserve_ratio_sysctl_handler,
2911 	},
2912 	{
2913 		.procname	= "drop_caches",
2914 		.data		= &sysctl_drop_caches,
2915 		.maxlen		= sizeof(int),
2916 		.mode		= 0200,
2917 		.proc_handler	= drop_caches_sysctl_handler,
2918 		.extra1		= SYSCTL_ONE,
2919 		.extra2		= &four,
2920 	},
2921 #ifdef CONFIG_COMPACTION
2922 	{
2923 		.procname	= "compact_memory",
2924 		.data		= &sysctl_compact_memory,
2925 		.maxlen		= sizeof(int),
2926 		.mode		= 0200,
2927 		.proc_handler	= sysctl_compaction_handler,
2928 	},
2929 	{
2930 		.procname	= "compaction_proactiveness",
2931 		.data		= &sysctl_compaction_proactiveness,
2932 		.maxlen		= sizeof(sysctl_compaction_proactiveness),
2933 		.mode		= 0644,
2934 		.proc_handler	= proc_dointvec_minmax,
2935 		.extra1		= SYSCTL_ZERO,
2936 		.extra2		= &one_hundred,
2937 	},
2938 	{
2939 		.procname	= "extfrag_threshold",
2940 		.data		= &sysctl_extfrag_threshold,
2941 		.maxlen		= sizeof(int),
2942 		.mode		= 0644,
2943 		.proc_handler	= proc_dointvec_minmax,
2944 		.extra1		= &min_extfrag_threshold,
2945 		.extra2		= &max_extfrag_threshold,
2946 	},
2947 	{
2948 		.procname	= "compact_unevictable_allowed",
2949 		.data		= &sysctl_compact_unevictable_allowed,
2950 		.maxlen		= sizeof(int),
2951 		.mode		= 0644,
2952 		.proc_handler	= proc_dointvec_minmax_warn_RT_change,
2953 		.extra1		= SYSCTL_ZERO,
2954 		.extra2		= SYSCTL_ONE,
2955 	},
2956 
2957 #endif /* CONFIG_COMPACTION */
2958 	{
2959 		.procname	= "min_free_kbytes",
2960 		.data		= &min_free_kbytes,
2961 		.maxlen		= sizeof(min_free_kbytes),
2962 		.mode		= 0644,
2963 		.proc_handler	= min_free_kbytes_sysctl_handler,
2964 		.extra1		= SYSCTL_ZERO,
2965 	},
2966 	{
2967 		.procname	= "watermark_boost_factor",
2968 		.data		= &watermark_boost_factor,
2969 		.maxlen		= sizeof(watermark_boost_factor),
2970 		.mode		= 0644,
2971 		.proc_handler	= proc_dointvec_minmax,
2972 		.extra1		= SYSCTL_ZERO,
2973 	},
2974 	{
2975 		.procname	= "watermark_scale_factor",
2976 		.data		= &watermark_scale_factor,
2977 		.maxlen		= sizeof(watermark_scale_factor),
2978 		.mode		= 0644,
2979 		.proc_handler	= watermark_scale_factor_sysctl_handler,
2980 		.extra1		= SYSCTL_ONE,
2981 		.extra2		= &one_thousand,
2982 	},
2983 	{
2984 		.procname	= "percpu_pagelist_fraction",
2985 		.data		= &percpu_pagelist_fraction,
2986 		.maxlen		= sizeof(percpu_pagelist_fraction),
2987 		.mode		= 0644,
2988 		.proc_handler	= percpu_pagelist_fraction_sysctl_handler,
2989 		.extra1		= SYSCTL_ZERO,
2990 	},
2991 	{
2992 		.procname	= "page_lock_unfairness",
2993 		.data		= &sysctl_page_lock_unfairness,
2994 		.maxlen		= sizeof(sysctl_page_lock_unfairness),
2995 		.mode		= 0644,
2996 		.proc_handler	= proc_dointvec_minmax,
2997 		.extra1		= SYSCTL_ZERO,
2998 	},
2999 #ifdef CONFIG_MMU
3000 	{
3001 		.procname	= "max_map_count",
3002 		.data		= &sysctl_max_map_count,
3003 		.maxlen		= sizeof(sysctl_max_map_count),
3004 		.mode		= 0644,
3005 		.proc_handler	= proc_dointvec_minmax,
3006 		.extra1		= SYSCTL_ZERO,
3007 	},
3008 #else
3009 	{
3010 		.procname	= "nr_trim_pages",
3011 		.data		= &sysctl_nr_trim_pages,
3012 		.maxlen		= sizeof(sysctl_nr_trim_pages),
3013 		.mode		= 0644,
3014 		.proc_handler	= proc_dointvec_minmax,
3015 		.extra1		= SYSCTL_ZERO,
3016 	},
3017 #endif
3018 	{
3019 		.procname	= "laptop_mode",
3020 		.data		= &laptop_mode,
3021 		.maxlen		= sizeof(laptop_mode),
3022 		.mode		= 0644,
3023 		.proc_handler	= proc_dointvec_jiffies,
3024 	},
3025 	{
3026 		.procname	= "block_dump",
3027 		.data		= &block_dump,
3028 		.maxlen		= sizeof(block_dump),
3029 		.mode		= 0644,
3030 		.proc_handler	= proc_dointvec_minmax,
3031 		.extra1		= SYSCTL_ZERO,
3032 	},
3033 	{
3034 		.procname	= "vfs_cache_pressure",
3035 		.data		= &sysctl_vfs_cache_pressure,
3036 		.maxlen		= sizeof(sysctl_vfs_cache_pressure),
3037 		.mode		= 0644,
3038 		.proc_handler	= proc_dointvec_minmax,
3039 		.extra1		= SYSCTL_ZERO,
3040 	},
3041 #if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
3042     defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
3043 	{
3044 		.procname	= "legacy_va_layout",
3045 		.data		= &sysctl_legacy_va_layout,
3046 		.maxlen		= sizeof(sysctl_legacy_va_layout),
3047 		.mode		= 0644,
3048 		.proc_handler	= proc_dointvec_minmax,
3049 		.extra1		= SYSCTL_ZERO,
3050 	},
3051 #endif
3052 #ifdef CONFIG_NUMA
3053 	{
3054 		.procname	= "zone_reclaim_mode",
3055 		.data		= &node_reclaim_mode,
3056 		.maxlen		= sizeof(node_reclaim_mode),
3057 		.mode		= 0644,
3058 		.proc_handler	= proc_dointvec_minmax,
3059 		.extra1		= SYSCTL_ZERO,
3060 	},
3061 	{
3062 		.procname	= "min_unmapped_ratio",
3063 		.data		= &sysctl_min_unmapped_ratio,
3064 		.maxlen		= sizeof(sysctl_min_unmapped_ratio),
3065 		.mode		= 0644,
3066 		.proc_handler	= sysctl_min_unmapped_ratio_sysctl_handler,
3067 		.extra1		= SYSCTL_ZERO,
3068 		.extra2		= &one_hundred,
3069 	},
3070 	{
3071 		.procname	= "min_slab_ratio",
3072 		.data		= &sysctl_min_slab_ratio,
3073 		.maxlen		= sizeof(sysctl_min_slab_ratio),
3074 		.mode		= 0644,
3075 		.proc_handler	= sysctl_min_slab_ratio_sysctl_handler,
3076 		.extra1		= SYSCTL_ZERO,
3077 		.extra2		= &one_hundred,
3078 	},
3079 #endif
3080 #ifdef CONFIG_SMP
3081 	{
3082 		.procname	= "stat_interval",
3083 		.data		= &sysctl_stat_interval,
3084 		.maxlen		= sizeof(sysctl_stat_interval),
3085 		.mode		= 0644,
3086 		.proc_handler	= proc_dointvec_jiffies,
3087 	},
3088 	{
3089 		.procname	= "stat_refresh",
3090 		.data		= NULL,
3091 		.maxlen		= 0,
3092 		.mode		= 0600,
3093 		.proc_handler	= vmstat_refresh,
3094 	},
3095 #endif
3096 #ifdef CONFIG_MMU
3097 	{
3098 		.procname	= "mmap_min_addr",
3099 		.data		= &dac_mmap_min_addr,
3100 		.maxlen		= sizeof(unsigned long),
3101 		.mode		= 0644,
3102 		.proc_handler	= mmap_min_addr_handler,
3103 	},
3104 #endif
3105 #ifdef CONFIG_NUMA
3106 	{
3107 		.procname	= "numa_zonelist_order",
3108 		.data		= &numa_zonelist_order,
3109 		.maxlen		= NUMA_ZONELIST_ORDER_LEN,
3110 		.mode		= 0644,
3111 		.proc_handler	= numa_zonelist_order_handler,
3112 	},
3113 #endif
3114 #if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
3115    (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
3116 	{
3117 		.procname	= "vdso_enabled",
3118 #ifdef CONFIG_X86_32
3119 		.data		= &vdso32_enabled,
3120 		.maxlen		= sizeof(vdso32_enabled),
3121 #else
3122 		.data		= &vdso_enabled,
3123 		.maxlen		= sizeof(vdso_enabled),
3124 #endif
3125 		.mode		= 0644,
3126 		.proc_handler	= proc_dointvec,
3127 		.extra1		= SYSCTL_ZERO,
3128 	},
3129 #endif
3130 #ifdef CONFIG_HIGHMEM
3131 	{
3132 		.procname	= "highmem_is_dirtyable",
3133 		.data		= &vm_highmem_is_dirtyable,
3134 		.maxlen		= sizeof(vm_highmem_is_dirtyable),
3135 		.mode		= 0644,
3136 		.proc_handler	= proc_dointvec_minmax,
3137 		.extra1		= SYSCTL_ZERO,
3138 		.extra2		= SYSCTL_ONE,
3139 	},
3140 #endif
3141 #ifdef CONFIG_MEMORY_FAILURE
3142 	{
3143 		.procname	= "memory_failure_early_kill",
3144 		.data		= &sysctl_memory_failure_early_kill,
3145 		.maxlen		= sizeof(sysctl_memory_failure_early_kill),
3146 		.mode		= 0644,
3147 		.proc_handler	= proc_dointvec_minmax,
3148 		.extra1		= SYSCTL_ZERO,
3149 		.extra2		= SYSCTL_ONE,
3150 	},
3151 	{
3152 		.procname	= "memory_failure_recovery",
3153 		.data		= &sysctl_memory_failure_recovery,
3154 		.maxlen		= sizeof(sysctl_memory_failure_recovery),
3155 		.mode		= 0644,
3156 		.proc_handler	= proc_dointvec_minmax,
3157 		.extra1		= SYSCTL_ZERO,
3158 		.extra2		= SYSCTL_ONE,
3159 	},
3160 #endif
3161 	{
3162 		.procname	= "user_reserve_kbytes",
3163 		.data		= &sysctl_user_reserve_kbytes,
3164 		.maxlen		= sizeof(sysctl_user_reserve_kbytes),
3165 		.mode		= 0644,
3166 		.proc_handler	= proc_doulongvec_minmax,
3167 	},
3168 	{
3169 		.procname	= "admin_reserve_kbytes",
3170 		.data		= &sysctl_admin_reserve_kbytes,
3171 		.maxlen		= sizeof(sysctl_admin_reserve_kbytes),
3172 		.mode		= 0644,
3173 		.proc_handler	= proc_doulongvec_minmax,
3174 	},
3175 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_BITS
3176 	{
3177 		.procname	= "mmap_rnd_bits",
3178 		.data		= &mmap_rnd_bits,
3179 		.maxlen		= sizeof(mmap_rnd_bits),
3180 		.mode		= 0600,
3181 		.proc_handler	= proc_dointvec_minmax,
3182 		.extra1		= (void *)&mmap_rnd_bits_min,
3183 		.extra2		= (void *)&mmap_rnd_bits_max,
3184 	},
3185 #endif
3186 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_COMPAT_BITS
3187 	{
3188 		.procname	= "mmap_rnd_compat_bits",
3189 		.data		= &mmap_rnd_compat_bits,
3190 		.maxlen		= sizeof(mmap_rnd_compat_bits),
3191 		.mode		= 0600,
3192 		.proc_handler	= proc_dointvec_minmax,
3193 		.extra1		= (void *)&mmap_rnd_compat_bits_min,
3194 		.extra2		= (void *)&mmap_rnd_compat_bits_max,
3195 	},
3196 #endif
3197 #ifdef CONFIG_USERFAULTFD
3198 	{
3199 		.procname	= "unprivileged_userfaultfd",
3200 		.data		= &sysctl_unprivileged_userfaultfd,
3201 		.maxlen		= sizeof(sysctl_unprivileged_userfaultfd),
3202 		.mode		= 0644,
3203 		.proc_handler	= proc_dointvec_minmax,
3204 		.extra1		= SYSCTL_ZERO,
3205 		.extra2		= SYSCTL_ONE,
3206 	},
3207 #endif
3208 	{ }
3209 };
3210 
3211 static struct ctl_table fs_table[] = {
3212 	{
3213 		.procname	= "inode-nr",
3214 		.data		= &inodes_stat,
3215 		.maxlen		= 2*sizeof(long),
3216 		.mode		= 0444,
3217 		.proc_handler	= proc_nr_inodes,
3218 	},
3219 	{
3220 		.procname	= "inode-state",
3221 		.data		= &inodes_stat,
3222 		.maxlen		= 7*sizeof(long),
3223 		.mode		= 0444,
3224 		.proc_handler	= proc_nr_inodes,
3225 	},
3226 	{
3227 		.procname	= "file-nr",
3228 		.data		= &files_stat,
3229 		.maxlen		= sizeof(files_stat),
3230 		.mode		= 0444,
3231 		.proc_handler	= proc_nr_files,
3232 	},
3233 	{
3234 		.procname	= "file-max",
3235 		.data		= &files_stat.max_files,
3236 		.maxlen		= sizeof(files_stat.max_files),
3237 		.mode		= 0644,
3238 		.proc_handler	= proc_doulongvec_minmax,
3239 		.extra1		= &zero_ul,
3240 		.extra2		= &long_max,
3241 	},
3242 	{
3243 		.procname	= "nr_open",
3244 		.data		= &sysctl_nr_open,
3245 		.maxlen		= sizeof(unsigned int),
3246 		.mode		= 0644,
3247 		.proc_handler	= proc_dointvec_minmax,
3248 		.extra1		= &sysctl_nr_open_min,
3249 		.extra2		= &sysctl_nr_open_max,
3250 	},
3251 	{
3252 		.procname	= "dentry-state",
3253 		.data		= &dentry_stat,
3254 		.maxlen		= 6*sizeof(long),
3255 		.mode		= 0444,
3256 		.proc_handler	= proc_nr_dentry,
3257 	},
3258 	{
3259 		.procname	= "overflowuid",
3260 		.data		= &fs_overflowuid,
3261 		.maxlen		= sizeof(int),
3262 		.mode		= 0644,
3263 		.proc_handler	= proc_dointvec_minmax,
3264 		.extra1		= &minolduid,
3265 		.extra2		= &maxolduid,
3266 	},
3267 	{
3268 		.procname	= "overflowgid",
3269 		.data		= &fs_overflowgid,
3270 		.maxlen		= sizeof(int),
3271 		.mode		= 0644,
3272 		.proc_handler	= proc_dointvec_minmax,
3273 		.extra1		= &minolduid,
3274 		.extra2		= &maxolduid,
3275 	},
3276 #ifdef CONFIG_FILE_LOCKING
3277 	{
3278 		.procname	= "leases-enable",
3279 		.data		= &leases_enable,
3280 		.maxlen		= sizeof(int),
3281 		.mode		= 0644,
3282 		.proc_handler	= proc_dointvec,
3283 	},
3284 #endif
3285 #ifdef CONFIG_DNOTIFY
3286 	{
3287 		.procname	= "dir-notify-enable",
3288 		.data		= &dir_notify_enable,
3289 		.maxlen		= sizeof(int),
3290 		.mode		= 0644,
3291 		.proc_handler	= proc_dointvec,
3292 	},
3293 #endif
3294 #ifdef CONFIG_MMU
3295 #ifdef CONFIG_FILE_LOCKING
3296 	{
3297 		.procname	= "lease-break-time",
3298 		.data		= &lease_break_time,
3299 		.maxlen		= sizeof(int),
3300 		.mode		= 0644,
3301 		.proc_handler	= proc_dointvec,
3302 	},
3303 #endif
3304 #ifdef CONFIG_AIO
3305 	{
3306 		.procname	= "aio-nr",
3307 		.data		= &aio_nr,
3308 		.maxlen		= sizeof(aio_nr),
3309 		.mode		= 0444,
3310 		.proc_handler	= proc_doulongvec_minmax,
3311 	},
3312 	{
3313 		.procname	= "aio-max-nr",
3314 		.data		= &aio_max_nr,
3315 		.maxlen		= sizeof(aio_max_nr),
3316 		.mode		= 0644,
3317 		.proc_handler	= proc_doulongvec_minmax,
3318 	},
3319 #endif /* CONFIG_AIO */
3320 #ifdef CONFIG_INOTIFY_USER
3321 	{
3322 		.procname	= "inotify",
3323 		.mode		= 0555,
3324 		.child		= inotify_table,
3325 	},
3326 #endif
3327 #ifdef CONFIG_EPOLL
3328 	{
3329 		.procname	= "epoll",
3330 		.mode		= 0555,
3331 		.child		= epoll_table,
3332 	},
3333 #endif
3334 #endif
3335 	{
3336 		.procname	= "protected_symlinks",
3337 		.data		= &sysctl_protected_symlinks,
3338 		.maxlen		= sizeof(int),
3339 		.mode		= 0600,
3340 		.proc_handler	= proc_dointvec_minmax,
3341 		.extra1		= SYSCTL_ZERO,
3342 		.extra2		= SYSCTL_ONE,
3343 	},
3344 	{
3345 		.procname	= "protected_hardlinks",
3346 		.data		= &sysctl_protected_hardlinks,
3347 		.maxlen		= sizeof(int),
3348 		.mode		= 0600,
3349 		.proc_handler	= proc_dointvec_minmax,
3350 		.extra1		= SYSCTL_ZERO,
3351 		.extra2		= SYSCTL_ONE,
3352 	},
3353 	{
3354 		.procname	= "protected_fifos",
3355 		.data		= &sysctl_protected_fifos,
3356 		.maxlen		= sizeof(int),
3357 		.mode		= 0600,
3358 		.proc_handler	= proc_dointvec_minmax,
3359 		.extra1		= SYSCTL_ZERO,
3360 		.extra2		= &two,
3361 	},
3362 	{
3363 		.procname	= "protected_regular",
3364 		.data		= &sysctl_protected_regular,
3365 		.maxlen		= sizeof(int),
3366 		.mode		= 0600,
3367 		.proc_handler	= proc_dointvec_minmax,
3368 		.extra1		= SYSCTL_ZERO,
3369 		.extra2		= &two,
3370 	},
3371 	{
3372 		.procname	= "suid_dumpable",
3373 		.data		= &suid_dumpable,
3374 		.maxlen		= sizeof(int),
3375 		.mode		= 0644,
3376 		.proc_handler	= proc_dointvec_minmax_coredump,
3377 		.extra1		= SYSCTL_ZERO,
3378 		.extra2		= &two,
3379 	},
3380 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
3381 	{
3382 		.procname	= "binfmt_misc",
3383 		.mode		= 0555,
3384 		.child		= sysctl_mount_point,
3385 	},
3386 #endif
3387 	{
3388 		.procname	= "pipe-max-size",
3389 		.data		= &pipe_max_size,
3390 		.maxlen		= sizeof(pipe_max_size),
3391 		.mode		= 0644,
3392 		.proc_handler	= proc_dopipe_max_size,
3393 	},
3394 	{
3395 		.procname	= "pipe-user-pages-hard",
3396 		.data		= &pipe_user_pages_hard,
3397 		.maxlen		= sizeof(pipe_user_pages_hard),
3398 		.mode		= 0644,
3399 		.proc_handler	= proc_doulongvec_minmax,
3400 	},
3401 	{
3402 		.procname	= "pipe-user-pages-soft",
3403 		.data		= &pipe_user_pages_soft,
3404 		.maxlen		= sizeof(pipe_user_pages_soft),
3405 		.mode		= 0644,
3406 		.proc_handler	= proc_doulongvec_minmax,
3407 	},
3408 	{
3409 		.procname	= "mount-max",
3410 		.data		= &sysctl_mount_max,
3411 		.maxlen		= sizeof(unsigned int),
3412 		.mode		= 0644,
3413 		.proc_handler	= proc_dointvec_minmax,
3414 		.extra1		= SYSCTL_ONE,
3415 	},
3416 	{ }
3417 };
3418 
3419 static struct ctl_table debug_table[] = {
3420 #ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
3421 	{
3422 		.procname	= "exception-trace",
3423 		.data		= &show_unhandled_signals,
3424 		.maxlen		= sizeof(int),
3425 		.mode		= 0644,
3426 		.proc_handler	= proc_dointvec
3427 	},
3428 #endif
3429 #if defined(CONFIG_OPTPROBES)
3430 	{
3431 		.procname	= "kprobes-optimization",
3432 		.data		= &sysctl_kprobes_optimization,
3433 		.maxlen		= sizeof(int),
3434 		.mode		= 0644,
3435 		.proc_handler	= proc_kprobes_optimization_handler,
3436 		.extra1		= SYSCTL_ZERO,
3437 		.extra2		= SYSCTL_ONE,
3438 	},
3439 #endif
3440 	{ }
3441 };
3442 
3443 static struct ctl_table dev_table[] = {
3444 	{ }
3445 };
3446 
3447 static struct ctl_table sysctl_base_table[] = {
3448 	{
3449 		.procname	= "kernel",
3450 		.mode		= 0555,
3451 		.child		= kern_table,
3452 	},
3453 	{
3454 		.procname	= "vm",
3455 		.mode		= 0555,
3456 		.child		= vm_table,
3457 	},
3458 	{
3459 		.procname	= "fs",
3460 		.mode		= 0555,
3461 		.child		= fs_table,
3462 	},
3463 	{
3464 		.procname	= "debug",
3465 		.mode		= 0555,
3466 		.child		= debug_table,
3467 	},
3468 	{
3469 		.procname	= "dev",
3470 		.mode		= 0555,
3471 		.child		= dev_table,
3472 	},
3473 	{ }
3474 };
3475 
3476 int __init sysctl_init(void)
3477 {
3478 	struct ctl_table_header *hdr;
3479 
3480 	hdr = register_sysctl_table(sysctl_base_table);
3481 	kmemleak_not_leak(hdr);
3482 	return 0;
3483 }
3484 #endif /* CONFIG_SYSCTL */
3485 /*
3486  * No sense putting this after each symbol definition, twice,
3487  * exception granted :-)
3488  */
3489 EXPORT_SYMBOL(proc_dointvec);
3490 EXPORT_SYMBOL(proc_douintvec);
3491 EXPORT_SYMBOL(proc_dointvec_jiffies);
3492 EXPORT_SYMBOL(proc_dointvec_minmax);
3493 EXPORT_SYMBOL_GPL(proc_douintvec_minmax);
3494 EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
3495 EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
3496 EXPORT_SYMBOL(proc_dostring);
3497 EXPORT_SYMBOL(proc_doulongvec_minmax);
3498 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
3499 EXPORT_SYMBOL(proc_do_large_bitmap);
3500