xref: /OK3568_Linux_fs/kernel/kernel/sysctl.c (revision 4882a59341e53eb6f0b4789bf948001014eff981)
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 /* External variables not in a header file. */
110 extern int extra_free_kbytes;
111 
112 /* Constants used for minimum and  maximum */
113 #ifdef CONFIG_LOCKUP_DETECTOR
114 static int sixty = 60;
115 #endif
116 
117 static int __maybe_unused neg_one = -1;
118 static int __maybe_unused two = 2;
119 static int __maybe_unused four = 4;
120 static unsigned long zero_ul;
121 static unsigned long one_ul = 1;
122 static unsigned long long_max = LONG_MAX;
123 static int one_hundred = 100;
124 static int two_hundred = 200;
125 static int one_thousand = 1000;
126 #ifdef CONFIG_PRINTK
127 static int ten_thousand = 10000;
128 #endif
129 #ifdef CONFIG_PERF_EVENTS
130 static int six_hundred_forty_kb = 640 * 1024;
131 #endif
132 
133 /* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
134 static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
135 
136 /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
137 static int maxolduid = 65535;
138 static int minolduid;
139 
140 static int ngroups_max = NGROUPS_MAX;
141 static const int cap_last_cap = CAP_LAST_CAP;
142 
143 /*
144  * This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
145  * and hung_task_check_interval_secs
146  */
147 #ifdef CONFIG_DETECT_HUNG_TASK
148 static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
149 #endif
150 
151 #ifdef CONFIG_INOTIFY_USER
152 #include <linux/inotify.h>
153 #endif
154 
155 #ifdef CONFIG_PROC_SYSCTL
156 
157 /**
158  * enum sysctl_writes_mode - supported sysctl write modes
159  *
160  * @SYSCTL_WRITES_LEGACY: each write syscall must fully contain the sysctl value
161  *	to be written, and multiple writes on the same sysctl file descriptor
162  *	will rewrite the sysctl value, regardless of file position. No warning
163  *	is issued when the initial position is not 0.
164  * @SYSCTL_WRITES_WARN: same as above but warn when the initial file position is
165  *	not 0.
166  * @SYSCTL_WRITES_STRICT: writes to numeric sysctl entries must always be at
167  *	file position 0 and the value must be fully contained in the buffer
168  *	sent to the write syscall. If dealing with strings respect the file
169  *	position, but restrict this to the max length of the buffer, anything
170  *	passed the max length will be ignored. Multiple writes will append
171  *	to the buffer.
172  *
173  * These write modes control how current file position affects the behavior of
174  * updating sysctl values through the proc interface on each write.
175  */
176 enum sysctl_writes_mode {
177 	SYSCTL_WRITES_LEGACY		= -1,
178 	SYSCTL_WRITES_WARN		= 0,
179 	SYSCTL_WRITES_STRICT		= 1,
180 };
181 
182 static enum sysctl_writes_mode sysctl_writes_strict = SYSCTL_WRITES_STRICT;
183 #endif /* CONFIG_PROC_SYSCTL */
184 
185 #if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
186     defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
187 int sysctl_legacy_va_layout;
188 #endif
189 
190 #ifdef CONFIG_SCHED_DEBUG
191 static int min_sched_granularity_ns = 100000;		/* 100 usecs */
192 static int max_sched_granularity_ns = NSEC_PER_SEC;	/* 1 second */
193 static int min_wakeup_granularity_ns;			/* 0 usecs */
194 static int max_wakeup_granularity_ns = NSEC_PER_SEC;	/* 1 second */
195 #ifdef CONFIG_SMP
196 static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
197 static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
198 #endif /* CONFIG_SMP */
199 #endif /* CONFIG_SCHED_DEBUG */
200 
201 #ifdef CONFIG_COMPACTION
202 static int min_extfrag_threshold;
203 static int max_extfrag_threshold = 1000;
204 #endif
205 
206 #endif /* CONFIG_SYSCTL */
207 
208 #if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_SYSCTL)
bpf_stats_handler(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)209 static int bpf_stats_handler(struct ctl_table *table, int write,
210 			     void *buffer, size_t *lenp, loff_t *ppos)
211 {
212 	struct static_key *key = (struct static_key *)table->data;
213 	static int saved_val;
214 	int val, ret;
215 	struct ctl_table tmp = {
216 		.data   = &val,
217 		.maxlen = sizeof(val),
218 		.mode   = table->mode,
219 		.extra1 = SYSCTL_ZERO,
220 		.extra2 = SYSCTL_ONE,
221 	};
222 
223 	if (write && !capable(CAP_SYS_ADMIN))
224 		return -EPERM;
225 
226 	mutex_lock(&bpf_stats_enabled_mutex);
227 	val = saved_val;
228 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
229 	if (write && !ret && val != saved_val) {
230 		if (val)
231 			static_key_slow_inc(key);
232 		else
233 			static_key_slow_dec(key);
234 		saved_val = val;
235 	}
236 	mutex_unlock(&bpf_stats_enabled_mutex);
237 	return ret;
238 }
239 
unpriv_ebpf_notify(int new_state)240 void __weak unpriv_ebpf_notify(int new_state)
241 {
242 }
243 
bpf_unpriv_handler(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)244 static int bpf_unpriv_handler(struct ctl_table *table, int write,
245 			      void *buffer, size_t *lenp, loff_t *ppos)
246 {
247 	int ret, unpriv_enable = *(int *)table->data;
248 	bool locked_state = unpriv_enable == 1;
249 	struct ctl_table tmp = *table;
250 
251 	if (write && !capable(CAP_SYS_ADMIN))
252 		return -EPERM;
253 
254 	tmp.data = &unpriv_enable;
255 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
256 	if (write && !ret) {
257 		if (locked_state && unpriv_enable != 1)
258 			return -EPERM;
259 		*(int *)table->data = unpriv_enable;
260 	}
261 
262 	unpriv_ebpf_notify(unpriv_enable);
263 
264 	return ret;
265 }
266 #endif /* CONFIG_BPF_SYSCALL && CONFIG_SYSCTL */
267 
268 /*
269  * /proc/sys support
270  */
271 
272 #ifdef CONFIG_PROC_SYSCTL
273 
_proc_do_string(char * data,int maxlen,int write,char * buffer,size_t * lenp,loff_t * ppos)274 static int _proc_do_string(char *data, int maxlen, int write,
275 		char *buffer, size_t *lenp, loff_t *ppos)
276 {
277 	size_t len;
278 	char c, *p;
279 
280 	if (!data || !maxlen || !*lenp) {
281 		*lenp = 0;
282 		return 0;
283 	}
284 
285 	if (write) {
286 		if (sysctl_writes_strict == SYSCTL_WRITES_STRICT) {
287 			/* Only continue writes not past the end of buffer. */
288 			len = strlen(data);
289 			if (len > maxlen - 1)
290 				len = maxlen - 1;
291 
292 			if (*ppos > len)
293 				return 0;
294 			len = *ppos;
295 		} else {
296 			/* Start writing from beginning of buffer. */
297 			len = 0;
298 		}
299 
300 		*ppos += *lenp;
301 		p = buffer;
302 		while ((p - buffer) < *lenp && len < maxlen - 1) {
303 			c = *(p++);
304 			if (c == 0 || c == '\n')
305 				break;
306 			data[len++] = c;
307 		}
308 		data[len] = 0;
309 	} else {
310 		len = strlen(data);
311 		if (len > maxlen)
312 			len = maxlen;
313 
314 		if (*ppos > len) {
315 			*lenp = 0;
316 			return 0;
317 		}
318 
319 		data += *ppos;
320 		len  -= *ppos;
321 
322 		if (len > *lenp)
323 			len = *lenp;
324 		if (len)
325 			memcpy(buffer, data, len);
326 		if (len < *lenp) {
327 			buffer[len] = '\n';
328 			len++;
329 		}
330 		*lenp = len;
331 		*ppos += len;
332 	}
333 	return 0;
334 }
335 
warn_sysctl_write(struct ctl_table * table)336 static void warn_sysctl_write(struct ctl_table *table)
337 {
338 	pr_warn_once("%s wrote to %s when file position was not 0!\n"
339 		"This will not be supported in the future. To silence this\n"
340 		"warning, set kernel.sysctl_writes_strict = -1\n",
341 		current->comm, table->procname);
342 }
343 
344 /**
345  * proc_first_pos_non_zero_ignore - check if first position is allowed
346  * @ppos: file position
347  * @table: the sysctl table
348  *
349  * Returns true if the first position is non-zero and the sysctl_writes_strict
350  * mode indicates this is not allowed for numeric input types. String proc
351  * handlers can ignore the return value.
352  */
proc_first_pos_non_zero_ignore(loff_t * ppos,struct ctl_table * table)353 static bool proc_first_pos_non_zero_ignore(loff_t *ppos,
354 					   struct ctl_table *table)
355 {
356 	if (!*ppos)
357 		return false;
358 
359 	switch (sysctl_writes_strict) {
360 	case SYSCTL_WRITES_STRICT:
361 		return true;
362 	case SYSCTL_WRITES_WARN:
363 		warn_sysctl_write(table);
364 		return false;
365 	default:
366 		return false;
367 	}
368 }
369 
370 /**
371  * proc_dostring - read a string sysctl
372  * @table: the sysctl table
373  * @write: %TRUE if this is a write to the sysctl file
374  * @buffer: the user buffer
375  * @lenp: the size of the user buffer
376  * @ppos: file position
377  *
378  * Reads/writes a string from/to the user buffer. If the kernel
379  * buffer provided is not large enough to hold the string, the
380  * string is truncated. The copied string is %NULL-terminated.
381  * If the string is being read by the user process, it is copied
382  * and a newline '\n' is added. It is truncated if the buffer is
383  * not large enough.
384  *
385  * Returns 0 on success.
386  */
proc_dostring(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)387 int proc_dostring(struct ctl_table *table, int write,
388 		  void *buffer, size_t *lenp, loff_t *ppos)
389 {
390 	if (write)
391 		proc_first_pos_non_zero_ignore(ppos, table);
392 
393 	return _proc_do_string(table->data, table->maxlen, write, buffer, lenp,
394 			ppos);
395 }
396 
proc_skip_spaces(char ** buf,size_t * size)397 static void proc_skip_spaces(char **buf, size_t *size)
398 {
399 	while (*size) {
400 		if (!isspace(**buf))
401 			break;
402 		(*size)--;
403 		(*buf)++;
404 	}
405 }
406 
proc_skip_char(char ** buf,size_t * size,const char v)407 static void proc_skip_char(char **buf, size_t *size, const char v)
408 {
409 	while (*size) {
410 		if (**buf != v)
411 			break;
412 		(*size)--;
413 		(*buf)++;
414 	}
415 }
416 
417 /**
418  * strtoul_lenient - parse an ASCII formatted integer from a buffer and only
419  *                   fail on overflow
420  *
421  * @cp: kernel buffer containing the string to parse
422  * @endp: pointer to store the trailing characters
423  * @base: the base to use
424  * @res: where the parsed integer will be stored
425  *
426  * In case of success 0 is returned and @res will contain the parsed integer,
427  * @endp will hold any trailing characters.
428  * This function will fail the parse on overflow. If there wasn't an overflow
429  * the function will defer the decision what characters count as invalid to the
430  * caller.
431  */
strtoul_lenient(const char * cp,char ** endp,unsigned int base,unsigned long * res)432 static int strtoul_lenient(const char *cp, char **endp, unsigned int base,
433 			   unsigned long *res)
434 {
435 	unsigned long long result;
436 	unsigned int rv;
437 
438 	cp = _parse_integer_fixup_radix(cp, &base);
439 	rv = _parse_integer(cp, base, &result);
440 	if ((rv & KSTRTOX_OVERFLOW) || (result != (unsigned long)result))
441 		return -ERANGE;
442 
443 	cp += rv;
444 
445 	if (endp)
446 		*endp = (char *)cp;
447 
448 	*res = (unsigned long)result;
449 	return 0;
450 }
451 
452 #define TMPBUFLEN 22
453 /**
454  * proc_get_long - reads an ASCII formatted integer from a user buffer
455  *
456  * @buf: a kernel buffer
457  * @size: size of the kernel buffer
458  * @val: this is where the number will be stored
459  * @neg: set to %TRUE if number is negative
460  * @perm_tr: a vector which contains the allowed trailers
461  * @perm_tr_len: size of the perm_tr vector
462  * @tr: pointer to store the trailer character
463  *
464  * In case of success %0 is returned and @buf and @size are updated with
465  * the amount of bytes read. If @tr is non-NULL and a trailing
466  * character exists (size is non-zero after returning from this
467  * function), @tr is updated with the trailing character.
468  */
proc_get_long(char ** buf,size_t * size,unsigned long * val,bool * neg,const char * perm_tr,unsigned perm_tr_len,char * tr)469 static int proc_get_long(char **buf, size_t *size,
470 			  unsigned long *val, bool *neg,
471 			  const char *perm_tr, unsigned perm_tr_len, char *tr)
472 {
473 	char *p, tmp[TMPBUFLEN];
474 	ssize_t len = *size;
475 
476 	if (len <= 0)
477 		return -EINVAL;
478 
479 	if (len > TMPBUFLEN - 1)
480 		len = TMPBUFLEN - 1;
481 
482 	memcpy(tmp, *buf, len);
483 
484 	tmp[len] = 0;
485 	p = tmp;
486 	if (*p == '-' && *size > 1) {
487 		*neg = true;
488 		p++;
489 	} else
490 		*neg = false;
491 	if (!isdigit(*p))
492 		return -EINVAL;
493 
494 	if (strtoul_lenient(p, &p, 0, val))
495 		return -EINVAL;
496 
497 	len = p - tmp;
498 
499 	/* We don't know if the next char is whitespace thus we may accept
500 	 * invalid integers (e.g. 1234...a) or two integers instead of one
501 	 * (e.g. 123...1). So lets not allow such large numbers. */
502 	if (len == TMPBUFLEN - 1)
503 		return -EINVAL;
504 
505 	if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
506 		return -EINVAL;
507 
508 	if (tr && (len < *size))
509 		*tr = *p;
510 
511 	*buf += len;
512 	*size -= len;
513 
514 	return 0;
515 }
516 
517 /**
518  * proc_put_long - converts an integer to a decimal ASCII formatted string
519  *
520  * @buf: the user buffer
521  * @size: the size of the user buffer
522  * @val: the integer to be converted
523  * @neg: sign of the number, %TRUE for negative
524  *
525  * In case of success @buf and @size are updated with the amount of bytes
526  * written.
527  */
proc_put_long(void ** buf,size_t * size,unsigned long val,bool neg)528 static void proc_put_long(void **buf, size_t *size, unsigned long val, bool neg)
529 {
530 	int len;
531 	char tmp[TMPBUFLEN], *p = tmp;
532 
533 	sprintf(p, "%s%lu", neg ? "-" : "", val);
534 	len = strlen(tmp);
535 	if (len > *size)
536 		len = *size;
537 	memcpy(*buf, tmp, len);
538 	*size -= len;
539 	*buf += len;
540 }
541 #undef TMPBUFLEN
542 
proc_put_char(void ** buf,size_t * size,char c)543 static void proc_put_char(void **buf, size_t *size, char c)
544 {
545 	if (*size) {
546 		char **buffer = (char **)buf;
547 		**buffer = c;
548 
549 		(*size)--;
550 		(*buffer)++;
551 		*buf = *buffer;
552 	}
553 }
554 
do_proc_dointvec_conv(bool * negp,unsigned long * lvalp,int * valp,int write,void * data)555 static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
556 				 int *valp,
557 				 int write, void *data)
558 {
559 	if (write) {
560 		if (*negp) {
561 			if (*lvalp > (unsigned long) INT_MAX + 1)
562 				return -EINVAL;
563 			WRITE_ONCE(*valp, -*lvalp);
564 		} else {
565 			if (*lvalp > (unsigned long) INT_MAX)
566 				return -EINVAL;
567 			WRITE_ONCE(*valp, *lvalp);
568 		}
569 	} else {
570 		int val = READ_ONCE(*valp);
571 		if (val < 0) {
572 			*negp = true;
573 			*lvalp = -(unsigned long)val;
574 		} else {
575 			*negp = false;
576 			*lvalp = (unsigned long)val;
577 		}
578 	}
579 	return 0;
580 }
581 
do_proc_douintvec_conv(unsigned long * lvalp,unsigned int * valp,int write,void * data)582 static int do_proc_douintvec_conv(unsigned long *lvalp,
583 				  unsigned int *valp,
584 				  int write, void *data)
585 {
586 	if (write) {
587 		if (*lvalp > UINT_MAX)
588 			return -EINVAL;
589 		WRITE_ONCE(*valp, *lvalp);
590 	} else {
591 		unsigned int val = READ_ONCE(*valp);
592 		*lvalp = (unsigned long)val;
593 	}
594 	return 0;
595 }
596 
597 static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
598 
__do_proc_dointvec(void * tbl_data,struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos,int (* conv)(bool * negp,unsigned long * lvalp,int * valp,int write,void * data),void * data)599 static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
600 		  int write, void *buffer,
601 		  size_t *lenp, loff_t *ppos,
602 		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
603 			      int write, void *data),
604 		  void *data)
605 {
606 	int *i, vleft, first = 1, err = 0;
607 	size_t left;
608 	char *p;
609 
610 	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
611 		*lenp = 0;
612 		return 0;
613 	}
614 
615 	i = (int *) tbl_data;
616 	vleft = table->maxlen / sizeof(*i);
617 	left = *lenp;
618 
619 	if (!conv)
620 		conv = do_proc_dointvec_conv;
621 
622 	if (write) {
623 		if (proc_first_pos_non_zero_ignore(ppos, table))
624 			goto out;
625 
626 		if (left > PAGE_SIZE - 1)
627 			left = PAGE_SIZE - 1;
628 		p = buffer;
629 	}
630 
631 	for (; left && vleft--; i++, first=0) {
632 		unsigned long lval;
633 		bool neg;
634 
635 		if (write) {
636 			proc_skip_spaces(&p, &left);
637 
638 			if (!left)
639 				break;
640 			err = proc_get_long(&p, &left, &lval, &neg,
641 					     proc_wspace_sep,
642 					     sizeof(proc_wspace_sep), NULL);
643 			if (err)
644 				break;
645 			if (conv(&neg, &lval, i, 1, data)) {
646 				err = -EINVAL;
647 				break;
648 			}
649 		} else {
650 			if (conv(&neg, &lval, i, 0, data)) {
651 				err = -EINVAL;
652 				break;
653 			}
654 			if (!first)
655 				proc_put_char(&buffer, &left, '\t');
656 			proc_put_long(&buffer, &left, lval, neg);
657 		}
658 	}
659 
660 	if (!write && !first && left && !err)
661 		proc_put_char(&buffer, &left, '\n');
662 	if (write && !err && left)
663 		proc_skip_spaces(&p, &left);
664 	if (write && first)
665 		return err ? : -EINVAL;
666 	*lenp -= left;
667 out:
668 	*ppos += *lenp;
669 	return err;
670 }
671 
do_proc_dointvec(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos,int (* conv)(bool * negp,unsigned long * lvalp,int * valp,int write,void * data),void * data)672 static int do_proc_dointvec(struct ctl_table *table, int write,
673 		  void *buffer, size_t *lenp, loff_t *ppos,
674 		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
675 			      int write, void *data),
676 		  void *data)
677 {
678 	return __do_proc_dointvec(table->data, table, write,
679 			buffer, lenp, ppos, conv, data);
680 }
681 
do_proc_douintvec_w(unsigned int * tbl_data,struct ctl_table * table,void * buffer,size_t * lenp,loff_t * ppos,int (* conv)(unsigned long * lvalp,unsigned int * valp,int write,void * data),void * data)682 static int do_proc_douintvec_w(unsigned int *tbl_data,
683 			       struct ctl_table *table,
684 			       void *buffer,
685 			       size_t *lenp, loff_t *ppos,
686 			       int (*conv)(unsigned long *lvalp,
687 					   unsigned int *valp,
688 					   int write, void *data),
689 			       void *data)
690 {
691 	unsigned long lval;
692 	int err = 0;
693 	size_t left;
694 	bool neg;
695 	char *p = buffer;
696 
697 	left = *lenp;
698 
699 	if (proc_first_pos_non_zero_ignore(ppos, table))
700 		goto bail_early;
701 
702 	if (left > PAGE_SIZE - 1)
703 		left = PAGE_SIZE - 1;
704 
705 	proc_skip_spaces(&p, &left);
706 	if (!left) {
707 		err = -EINVAL;
708 		goto out_free;
709 	}
710 
711 	err = proc_get_long(&p, &left, &lval, &neg,
712 			     proc_wspace_sep,
713 			     sizeof(proc_wspace_sep), NULL);
714 	if (err || neg) {
715 		err = -EINVAL;
716 		goto out_free;
717 	}
718 
719 	if (conv(&lval, tbl_data, 1, data)) {
720 		err = -EINVAL;
721 		goto out_free;
722 	}
723 
724 	if (!err && left)
725 		proc_skip_spaces(&p, &left);
726 
727 out_free:
728 	if (err)
729 		return -EINVAL;
730 
731 	return 0;
732 
733 	/* This is in keeping with old __do_proc_dointvec() */
734 bail_early:
735 	*ppos += *lenp;
736 	return err;
737 }
738 
do_proc_douintvec_r(unsigned int * tbl_data,void * buffer,size_t * lenp,loff_t * ppos,int (* conv)(unsigned long * lvalp,unsigned int * valp,int write,void * data),void * data)739 static int do_proc_douintvec_r(unsigned int *tbl_data, void *buffer,
740 			       size_t *lenp, loff_t *ppos,
741 			       int (*conv)(unsigned long *lvalp,
742 					   unsigned int *valp,
743 					   int write, void *data),
744 			       void *data)
745 {
746 	unsigned long lval;
747 	int err = 0;
748 	size_t left;
749 
750 	left = *lenp;
751 
752 	if (conv(&lval, tbl_data, 0, data)) {
753 		err = -EINVAL;
754 		goto out;
755 	}
756 
757 	proc_put_long(&buffer, &left, lval, false);
758 	if (!left)
759 		goto out;
760 
761 	proc_put_char(&buffer, &left, '\n');
762 
763 out:
764 	*lenp -= left;
765 	*ppos += *lenp;
766 
767 	return err;
768 }
769 
__do_proc_douintvec(void * tbl_data,struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos,int (* conv)(unsigned long * lvalp,unsigned int * valp,int write,void * data),void * data)770 static int __do_proc_douintvec(void *tbl_data, struct ctl_table *table,
771 			       int write, void *buffer,
772 			       size_t *lenp, loff_t *ppos,
773 			       int (*conv)(unsigned long *lvalp,
774 					   unsigned int *valp,
775 					   int write, void *data),
776 			       void *data)
777 {
778 	unsigned int *i, vleft;
779 
780 	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
781 		*lenp = 0;
782 		return 0;
783 	}
784 
785 	i = (unsigned int *) tbl_data;
786 	vleft = table->maxlen / sizeof(*i);
787 
788 	/*
789 	 * Arrays are not supported, keep this simple. *Do not* add
790 	 * support for them.
791 	 */
792 	if (vleft != 1) {
793 		*lenp = 0;
794 		return -EINVAL;
795 	}
796 
797 	if (!conv)
798 		conv = do_proc_douintvec_conv;
799 
800 	if (write)
801 		return do_proc_douintvec_w(i, table, buffer, lenp, ppos,
802 					   conv, data);
803 	return do_proc_douintvec_r(i, buffer, lenp, ppos, conv, data);
804 }
805 
do_proc_douintvec(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos,int (* conv)(unsigned long * lvalp,unsigned int * valp,int write,void * data),void * data)806 static int do_proc_douintvec(struct ctl_table *table, int write,
807 			     void *buffer, size_t *lenp, loff_t *ppos,
808 			     int (*conv)(unsigned long *lvalp,
809 					 unsigned int *valp,
810 					 int write, void *data),
811 			     void *data)
812 {
813 	return __do_proc_douintvec(table->data, table, write,
814 				   buffer, lenp, ppos, conv, data);
815 }
816 
817 /**
818  * proc_dointvec - read a vector of integers
819  * @table: the sysctl table
820  * @write: %TRUE if this is a write to the sysctl file
821  * @buffer: the user buffer
822  * @lenp: the size of the user buffer
823  * @ppos: file position
824  *
825  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
826  * values from/to the user buffer, treated as an ASCII string.
827  *
828  * Returns 0 on success.
829  */
proc_dointvec(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)830 int proc_dointvec(struct ctl_table *table, int write, void *buffer,
831 		  size_t *lenp, loff_t *ppos)
832 {
833 	return do_proc_dointvec(table, write, buffer, lenp, ppos, NULL, NULL);
834 }
835 
836 #ifdef CONFIG_COMPACTION
proc_dointvec_minmax_warn_RT_change(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)837 static int proc_dointvec_minmax_warn_RT_change(struct ctl_table *table,
838 		int write, void *buffer, size_t *lenp, loff_t *ppos)
839 {
840 	int ret, old;
841 
842 	if (!IS_ENABLED(CONFIG_PREEMPT_RT) || !write)
843 		return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
844 
845 	old = *(int *)table->data;
846 	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
847 	if (ret)
848 		return ret;
849 	if (old != *(int *)table->data)
850 		pr_warn_once("sysctl attribute %s changed by %s[%d]\n",
851 			     table->procname, current->comm,
852 			     task_pid_nr(current));
853 	return ret;
854 }
855 #endif
856 
857 /**
858  * proc_douintvec - read a vector of unsigned integers
859  * @table: the sysctl table
860  * @write: %TRUE if this is a write to the sysctl file
861  * @buffer: the user buffer
862  * @lenp: the size of the user buffer
863  * @ppos: file position
864  *
865  * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
866  * values from/to the user buffer, treated as an ASCII string.
867  *
868  * Returns 0 on success.
869  */
proc_douintvec(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)870 int proc_douintvec(struct ctl_table *table, int write, void *buffer,
871 		size_t *lenp, loff_t *ppos)
872 {
873 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
874 				 do_proc_douintvec_conv, NULL);
875 }
876 
877 /*
878  * Taint values can only be increased
879  * This means we can safely use a temporary.
880  */
proc_taint(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)881 static int proc_taint(struct ctl_table *table, int write,
882 			       void *buffer, size_t *lenp, loff_t *ppos)
883 {
884 	struct ctl_table t;
885 	unsigned long tmptaint = get_taint();
886 	int err;
887 
888 	if (write && !capable(CAP_SYS_ADMIN))
889 		return -EPERM;
890 
891 	t = *table;
892 	t.data = &tmptaint;
893 	err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
894 	if (err < 0)
895 		return err;
896 
897 	if (write) {
898 		int i;
899 
900 		/*
901 		 * If we are relying on panic_on_taint not producing
902 		 * false positives due to userspace input, bail out
903 		 * before setting the requested taint flags.
904 		 */
905 		if (panic_on_taint_nousertaint && (tmptaint & panic_on_taint))
906 			return -EINVAL;
907 
908 		/*
909 		 * Poor man's atomic or. Not worth adding a primitive
910 		 * to everyone's atomic.h for this
911 		 */
912 		for (i = 0; i < TAINT_FLAGS_COUNT; i++)
913 			if ((1UL << i) & tmptaint)
914 				add_taint(i, LOCKDEP_STILL_OK);
915 	}
916 
917 	return err;
918 }
919 
920 #ifdef CONFIG_PRINTK
proc_dointvec_minmax_sysadmin(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)921 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
922 				void *buffer, size_t *lenp, loff_t *ppos)
923 {
924 	if (write && !capable(CAP_SYS_ADMIN))
925 		return -EPERM;
926 
927 	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
928 }
929 #endif
930 
931 /**
932  * struct do_proc_dointvec_minmax_conv_param - proc_dointvec_minmax() range checking structure
933  * @min: pointer to minimum allowable value
934  * @max: pointer to maximum allowable value
935  *
936  * The do_proc_dointvec_minmax_conv_param structure provides the
937  * minimum and maximum values for doing range checking for those sysctl
938  * parameters that use the proc_dointvec_minmax() handler.
939  */
940 struct do_proc_dointvec_minmax_conv_param {
941 	int *min;
942 	int *max;
943 };
944 
do_proc_dointvec_minmax_conv(bool * negp,unsigned long * lvalp,int * valp,int write,void * data)945 static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
946 					int *valp,
947 					int write, void *data)
948 {
949 	int tmp, ret;
950 	struct do_proc_dointvec_minmax_conv_param *param = data;
951 	/*
952 	 * If writing, first do so via a temporary local int so we can
953 	 * bounds-check it before touching *valp.
954 	 */
955 	int *ip = write ? &tmp : valp;
956 
957 	ret = do_proc_dointvec_conv(negp, lvalp, ip, write, data);
958 	if (ret)
959 		return ret;
960 
961 	if (write) {
962 		if ((param->min && *param->min > tmp) ||
963 		    (param->max && *param->max < tmp))
964 			return -EINVAL;
965 		WRITE_ONCE(*valp, tmp);
966 	}
967 
968 	return 0;
969 }
970 
971 /**
972  * proc_dointvec_minmax - read a vector of integers with min/max values
973  * @table: the sysctl table
974  * @write: %TRUE if this is a write to the sysctl file
975  * @buffer: the user buffer
976  * @lenp: the size of the user buffer
977  * @ppos: file position
978  *
979  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
980  * values from/to the user buffer, treated as an ASCII string.
981  *
982  * This routine will ensure the values are within the range specified by
983  * table->extra1 (min) and table->extra2 (max).
984  *
985  * Returns 0 on success or -EINVAL on write when the range check fails.
986  */
proc_dointvec_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)987 int proc_dointvec_minmax(struct ctl_table *table, int write,
988 		  void *buffer, size_t *lenp, loff_t *ppos)
989 {
990 	struct do_proc_dointvec_minmax_conv_param param = {
991 		.min = (int *) table->extra1,
992 		.max = (int *) table->extra2,
993 	};
994 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
995 				do_proc_dointvec_minmax_conv, &param);
996 }
997 
998 /**
999  * struct do_proc_douintvec_minmax_conv_param - proc_douintvec_minmax() range checking structure
1000  * @min: pointer to minimum allowable value
1001  * @max: pointer to maximum allowable value
1002  *
1003  * The do_proc_douintvec_minmax_conv_param structure provides the
1004  * minimum and maximum values for doing range checking for those sysctl
1005  * parameters that use the proc_douintvec_minmax() handler.
1006  */
1007 struct do_proc_douintvec_minmax_conv_param {
1008 	unsigned int *min;
1009 	unsigned int *max;
1010 };
1011 
do_proc_douintvec_minmax_conv(unsigned long * lvalp,unsigned int * valp,int write,void * data)1012 static int do_proc_douintvec_minmax_conv(unsigned long *lvalp,
1013 					 unsigned int *valp,
1014 					 int write, void *data)
1015 {
1016 	int ret;
1017 	unsigned int tmp;
1018 	struct do_proc_douintvec_minmax_conv_param *param = data;
1019 	/* write via temporary local uint for bounds-checking */
1020 	unsigned int *up = write ? &tmp : valp;
1021 
1022 	ret = do_proc_douintvec_conv(lvalp, up, write, data);
1023 	if (ret)
1024 		return ret;
1025 
1026 	if (write) {
1027 		if ((param->min && *param->min > tmp) ||
1028 		    (param->max && *param->max < tmp))
1029 			return -ERANGE;
1030 
1031 		WRITE_ONCE(*valp, tmp);
1032 	}
1033 
1034 	return 0;
1035 }
1036 
1037 /**
1038  * proc_douintvec_minmax - read a vector of unsigned ints 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(unsigned int) unsigned integer
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). There is a final sanity
1051  * check for UINT_MAX to avoid having to support wrap around uses from
1052  * userspace.
1053  *
1054  * Returns 0 on success or -ERANGE on write when the range check fails.
1055  */
proc_douintvec_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1056 int proc_douintvec_minmax(struct ctl_table *table, int write,
1057 			  void *buffer, size_t *lenp, loff_t *ppos)
1058 {
1059 	struct do_proc_douintvec_minmax_conv_param param = {
1060 		.min = (unsigned int *) table->extra1,
1061 		.max = (unsigned int *) table->extra2,
1062 	};
1063 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1064 				 do_proc_douintvec_minmax_conv, &param);
1065 }
1066 
do_proc_dopipe_max_size_conv(unsigned long * lvalp,unsigned int * valp,int write,void * data)1067 static int do_proc_dopipe_max_size_conv(unsigned long *lvalp,
1068 					unsigned int *valp,
1069 					int write, void *data)
1070 {
1071 	if (write) {
1072 		unsigned int val;
1073 
1074 		val = round_pipe_size(*lvalp);
1075 		if (val == 0)
1076 			return -EINVAL;
1077 
1078 		*valp = val;
1079 	} else {
1080 		unsigned int val = *valp;
1081 		*lvalp = (unsigned long) val;
1082 	}
1083 
1084 	return 0;
1085 }
1086 
proc_dopipe_max_size(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1087 static int proc_dopipe_max_size(struct ctl_table *table, int write,
1088 				void *buffer, size_t *lenp, loff_t *ppos)
1089 {
1090 	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1091 				 do_proc_dopipe_max_size_conv, NULL);
1092 }
1093 
validate_coredump_safety(void)1094 static void validate_coredump_safety(void)
1095 {
1096 #ifdef CONFIG_COREDUMP
1097 	if (suid_dumpable == SUID_DUMP_ROOT &&
1098 	    core_pattern[0] != '/' && core_pattern[0] != '|') {
1099 		printk(KERN_WARNING
1100 "Unsafe core_pattern used with fs.suid_dumpable=2.\n"
1101 "Pipe handler or fully qualified core dump path required.\n"
1102 "Set kernel.core_pattern before fs.suid_dumpable.\n"
1103 		);
1104 	}
1105 #endif
1106 }
1107 
proc_dointvec_minmax_coredump(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1108 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
1109 		void *buffer, size_t *lenp, loff_t *ppos)
1110 {
1111 	int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
1112 	if (!error)
1113 		validate_coredump_safety();
1114 	return error;
1115 }
1116 
1117 #ifdef CONFIG_COREDUMP
proc_dostring_coredump(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1118 static int proc_dostring_coredump(struct ctl_table *table, int write,
1119 		  void *buffer, size_t *lenp, loff_t *ppos)
1120 {
1121 	int error = proc_dostring(table, write, buffer, lenp, ppos);
1122 	if (!error)
1123 		validate_coredump_safety();
1124 	return error;
1125 }
1126 #endif
1127 
1128 #ifdef CONFIG_MAGIC_SYSRQ
sysrq_sysctl_handler(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1129 static int sysrq_sysctl_handler(struct ctl_table *table, int write,
1130 				void *buffer, size_t *lenp, loff_t *ppos)
1131 {
1132 	int tmp, ret;
1133 
1134 	tmp = sysrq_mask();
1135 
1136 	ret = __do_proc_dointvec(&tmp, table, write, buffer,
1137 			       lenp, ppos, NULL, NULL);
1138 	if (ret || !write)
1139 		return ret;
1140 
1141 	if (write)
1142 		sysrq_toggle_support(tmp);
1143 
1144 	return 0;
1145 }
1146 #endif
1147 
__do_proc_doulongvec_minmax(void * data,struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos,unsigned long convmul,unsigned long convdiv)1148 static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table,
1149 		int write, void *buffer, size_t *lenp, loff_t *ppos,
1150 		unsigned long convmul, unsigned long convdiv)
1151 {
1152 	unsigned long *i, *min, *max;
1153 	int vleft, first = 1, err = 0;
1154 	size_t left;
1155 	char *p;
1156 
1157 	if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
1158 		*lenp = 0;
1159 		return 0;
1160 	}
1161 
1162 	i = (unsigned long *) data;
1163 	min = (unsigned long *) table->extra1;
1164 	max = (unsigned long *) table->extra2;
1165 	vleft = table->maxlen / sizeof(unsigned long);
1166 	left = *lenp;
1167 
1168 	if (write) {
1169 		if (proc_first_pos_non_zero_ignore(ppos, table))
1170 			goto out;
1171 
1172 		if (left > PAGE_SIZE - 1)
1173 			left = PAGE_SIZE - 1;
1174 		p = buffer;
1175 	}
1176 
1177 	for (; left && vleft--; i++, first = 0) {
1178 		unsigned long val;
1179 
1180 		if (write) {
1181 			bool neg;
1182 
1183 			proc_skip_spaces(&p, &left);
1184 			if (!left)
1185 				break;
1186 
1187 			err = proc_get_long(&p, &left, &val, &neg,
1188 					     proc_wspace_sep,
1189 					     sizeof(proc_wspace_sep), NULL);
1190 			if (err)
1191 				break;
1192 			if (neg)
1193 				continue;
1194 			val = convmul * val / convdiv;
1195 			if ((min && val < *min) || (max && val > *max)) {
1196 				err = -EINVAL;
1197 				break;
1198 			}
1199 			WRITE_ONCE(*i, val);
1200 		} else {
1201 			val = convdiv * READ_ONCE(*i) / convmul;
1202 			if (!first)
1203 				proc_put_char(&buffer, &left, '\t');
1204 			proc_put_long(&buffer, &left, val, false);
1205 		}
1206 	}
1207 
1208 	if (!write && !first && left && !err)
1209 		proc_put_char(&buffer, &left, '\n');
1210 	if (write && !err)
1211 		proc_skip_spaces(&p, &left);
1212 	if (write && first)
1213 		return err ? : -EINVAL;
1214 	*lenp -= left;
1215 out:
1216 	*ppos += *lenp;
1217 	return err;
1218 }
1219 
do_proc_doulongvec_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos,unsigned long convmul,unsigned long convdiv)1220 static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
1221 		void *buffer, size_t *lenp, loff_t *ppos, unsigned long convmul,
1222 		unsigned long convdiv)
1223 {
1224 	return __do_proc_doulongvec_minmax(table->data, table, write,
1225 			buffer, lenp, ppos, convmul, convdiv);
1226 }
1227 
1228 /**
1229  * proc_doulongvec_minmax - read a vector of long integers with min/max values
1230  * @table: the sysctl table
1231  * @write: %TRUE if this is a write to the sysctl file
1232  * @buffer: the user buffer
1233  * @lenp: the size of the user buffer
1234  * @ppos: file position
1235  *
1236  * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1237  * values from/to the user buffer, treated as an ASCII string.
1238  *
1239  * This routine will ensure the values are within the range specified by
1240  * table->extra1 (min) and table->extra2 (max).
1241  *
1242  * Returns 0 on success.
1243  */
proc_doulongvec_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1244 int proc_doulongvec_minmax(struct ctl_table *table, int write,
1245 			   void *buffer, size_t *lenp, loff_t *ppos)
1246 {
1247     return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
1248 }
1249 
1250 /**
1251  * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
1252  * @table: the sysctl table
1253  * @write: %TRUE if this is a write to the sysctl file
1254  * @buffer: the user buffer
1255  * @lenp: the size of the user buffer
1256  * @ppos: file position
1257  *
1258  * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1259  * values from/to the user buffer, treated as an ASCII string. The values
1260  * are treated as milliseconds, and converted to jiffies when they are stored.
1261  *
1262  * This routine will ensure the values are within the range specified by
1263  * table->extra1 (min) and table->extra2 (max).
1264  *
1265  * Returns 0 on success.
1266  */
proc_doulongvec_ms_jiffies_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1267 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1268 				      void *buffer, size_t *lenp, loff_t *ppos)
1269 {
1270     return do_proc_doulongvec_minmax(table, write, buffer,
1271 				     lenp, ppos, HZ, 1000l);
1272 }
1273 
1274 
do_proc_dointvec_jiffies_conv(bool * negp,unsigned long * lvalp,int * valp,int write,void * data)1275 static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
1276 					 int *valp,
1277 					 int write, void *data)
1278 {
1279 	if (write) {
1280 		if (*lvalp > INT_MAX / HZ)
1281 			return 1;
1282 		if (*negp)
1283 			WRITE_ONCE(*valp, -*lvalp * HZ);
1284 		else
1285 			WRITE_ONCE(*valp, *lvalp * HZ);
1286 	} else {
1287 		int val = READ_ONCE(*valp);
1288 		unsigned long lval;
1289 		if (val < 0) {
1290 			*negp = true;
1291 			lval = -(unsigned long)val;
1292 		} else {
1293 			*negp = false;
1294 			lval = (unsigned long)val;
1295 		}
1296 		*lvalp = lval / HZ;
1297 	}
1298 	return 0;
1299 }
1300 
do_proc_dointvec_userhz_jiffies_conv(bool * negp,unsigned long * lvalp,int * valp,int write,void * data)1301 static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
1302 						int *valp,
1303 						int write, void *data)
1304 {
1305 	if (write) {
1306 		if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
1307 			return 1;
1308 		*valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
1309 	} else {
1310 		int val = *valp;
1311 		unsigned long lval;
1312 		if (val < 0) {
1313 			*negp = true;
1314 			lval = -(unsigned long)val;
1315 		} else {
1316 			*negp = false;
1317 			lval = (unsigned long)val;
1318 		}
1319 		*lvalp = jiffies_to_clock_t(lval);
1320 	}
1321 	return 0;
1322 }
1323 
do_proc_dointvec_ms_jiffies_conv(bool * negp,unsigned long * lvalp,int * valp,int write,void * data)1324 static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
1325 					    int *valp,
1326 					    int write, void *data)
1327 {
1328 	if (write) {
1329 		unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
1330 
1331 		if (jif > INT_MAX)
1332 			return 1;
1333 		WRITE_ONCE(*valp, (int)jif);
1334 	} else {
1335 		int val = READ_ONCE(*valp);
1336 		unsigned long lval;
1337 		if (val < 0) {
1338 			*negp = true;
1339 			lval = -(unsigned long)val;
1340 		} else {
1341 			*negp = false;
1342 			lval = (unsigned long)val;
1343 		}
1344 		*lvalp = jiffies_to_msecs(lval);
1345 	}
1346 	return 0;
1347 }
1348 
1349 /**
1350  * proc_dointvec_jiffies - read a vector of integers as seconds
1351  * @table: the sysctl table
1352  * @write: %TRUE if this is a write to the sysctl file
1353  * @buffer: the user buffer
1354  * @lenp: the size of the user buffer
1355  * @ppos: file position
1356  *
1357  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1358  * values from/to the user buffer, treated as an ASCII string.
1359  * The values read are assumed to be in seconds, and are converted into
1360  * jiffies.
1361  *
1362  * Returns 0 on success.
1363  */
proc_dointvec_jiffies(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1364 int proc_dointvec_jiffies(struct ctl_table *table, int write,
1365 			  void *buffer, size_t *lenp, loff_t *ppos)
1366 {
1367     return do_proc_dointvec(table,write,buffer,lenp,ppos,
1368 		    	    do_proc_dointvec_jiffies_conv,NULL);
1369 }
1370 
1371 /**
1372  * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
1373  * @table: the sysctl table
1374  * @write: %TRUE if this is a write to the sysctl file
1375  * @buffer: the user buffer
1376  * @lenp: the size of the user buffer
1377  * @ppos: pointer to the file position
1378  *
1379  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1380  * values from/to the user buffer, treated as an ASCII string.
1381  * The values read are assumed to be in 1/USER_HZ seconds, and
1382  * are converted into jiffies.
1383  *
1384  * Returns 0 on success.
1385  */
proc_dointvec_userhz_jiffies(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1386 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1387 				 void *buffer, size_t *lenp, loff_t *ppos)
1388 {
1389     return do_proc_dointvec(table,write,buffer,lenp,ppos,
1390 		    	    do_proc_dointvec_userhz_jiffies_conv,NULL);
1391 }
1392 
1393 /**
1394  * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
1395  * @table: the sysctl table
1396  * @write: %TRUE if this is a write to the sysctl file
1397  * @buffer: the user buffer
1398  * @lenp: the size of the user buffer
1399  * @ppos: file position
1400  * @ppos: the current position in the file
1401  *
1402  * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1403  * values from/to the user buffer, treated as an ASCII string.
1404  * The values read are assumed to be in 1/1000 seconds, and
1405  * are converted into jiffies.
1406  *
1407  * Returns 0 on success.
1408  */
proc_dointvec_ms_jiffies(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1409 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write, void *buffer,
1410 		size_t *lenp, loff_t *ppos)
1411 {
1412 	return do_proc_dointvec(table, write, buffer, lenp, ppos,
1413 				do_proc_dointvec_ms_jiffies_conv, NULL);
1414 }
1415 
proc_do_cad_pid(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1416 static int proc_do_cad_pid(struct ctl_table *table, int write, void *buffer,
1417 		size_t *lenp, loff_t *ppos)
1418 {
1419 	struct pid *new_pid;
1420 	pid_t tmp;
1421 	int r;
1422 
1423 	tmp = pid_vnr(cad_pid);
1424 
1425 	r = __do_proc_dointvec(&tmp, table, write, buffer,
1426 			       lenp, ppos, NULL, NULL);
1427 	if (r || !write)
1428 		return r;
1429 
1430 	new_pid = find_get_pid(tmp);
1431 	if (!new_pid)
1432 		return -ESRCH;
1433 
1434 	put_pid(xchg(&cad_pid, new_pid));
1435 	return 0;
1436 }
1437 
1438 /**
1439  * proc_do_large_bitmap - read/write from/to a large bitmap
1440  * @table: the sysctl table
1441  * @write: %TRUE if this is a write to the sysctl file
1442  * @buffer: the user buffer
1443  * @lenp: the size of the user buffer
1444  * @ppos: file position
1445  *
1446  * The bitmap is stored at table->data and the bitmap length (in bits)
1447  * in table->maxlen.
1448  *
1449  * We use a range comma separated format (e.g. 1,3-4,10-10) so that
1450  * large bitmaps may be represented in a compact manner. Writing into
1451  * the file will clear the bitmap then update it with the given input.
1452  *
1453  * Returns 0 on success.
1454  */
proc_do_large_bitmap(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1455 int proc_do_large_bitmap(struct ctl_table *table, int write,
1456 			 void *buffer, size_t *lenp, loff_t *ppos)
1457 {
1458 	int err = 0;
1459 	bool first = 1;
1460 	size_t left = *lenp;
1461 	unsigned long bitmap_len = table->maxlen;
1462 	unsigned long *bitmap = *(unsigned long **) table->data;
1463 	unsigned long *tmp_bitmap = NULL;
1464 	char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
1465 
1466 	if (!bitmap || !bitmap_len || !left || (*ppos && !write)) {
1467 		*lenp = 0;
1468 		return 0;
1469 	}
1470 
1471 	if (write) {
1472 		char *p = buffer;
1473 		size_t skipped = 0;
1474 
1475 		if (left > PAGE_SIZE - 1) {
1476 			left = PAGE_SIZE - 1;
1477 			/* How much of the buffer we'll skip this pass */
1478 			skipped = *lenp - left;
1479 		}
1480 
1481 		tmp_bitmap = bitmap_zalloc(bitmap_len, GFP_KERNEL);
1482 		if (!tmp_bitmap)
1483 			return -ENOMEM;
1484 		proc_skip_char(&p, &left, '\n');
1485 		while (!err && left) {
1486 			unsigned long val_a, val_b;
1487 			bool neg;
1488 			size_t saved_left;
1489 
1490 			/* In case we stop parsing mid-number, we can reset */
1491 			saved_left = left;
1492 			err = proc_get_long(&p, &left, &val_a, &neg, tr_a,
1493 					     sizeof(tr_a), &c);
1494 			/*
1495 			 * If we consumed the entirety of a truncated buffer or
1496 			 * only one char is left (may be a "-"), then stop here,
1497 			 * reset, & come back for more.
1498 			 */
1499 			if ((left <= 1) && skipped) {
1500 				left = saved_left;
1501 				break;
1502 			}
1503 
1504 			if (err)
1505 				break;
1506 			if (val_a >= bitmap_len || neg) {
1507 				err = -EINVAL;
1508 				break;
1509 			}
1510 
1511 			val_b = val_a;
1512 			if (left) {
1513 				p++;
1514 				left--;
1515 			}
1516 
1517 			if (c == '-') {
1518 				err = proc_get_long(&p, &left, &val_b,
1519 						     &neg, tr_b, sizeof(tr_b),
1520 						     &c);
1521 				/*
1522 				 * If we consumed all of a truncated buffer or
1523 				 * then stop here, reset, & come back for more.
1524 				 */
1525 				if (!left && skipped) {
1526 					left = saved_left;
1527 					break;
1528 				}
1529 
1530 				if (err)
1531 					break;
1532 				if (val_b >= bitmap_len || neg ||
1533 				    val_a > val_b) {
1534 					err = -EINVAL;
1535 					break;
1536 				}
1537 				if (left) {
1538 					p++;
1539 					left--;
1540 				}
1541 			}
1542 
1543 			bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
1544 			first = 0;
1545 			proc_skip_char(&p, &left, '\n');
1546 		}
1547 		left += skipped;
1548 	} else {
1549 		unsigned long bit_a, bit_b = 0;
1550 
1551 		while (left) {
1552 			bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
1553 			if (bit_a >= bitmap_len)
1554 				break;
1555 			bit_b = find_next_zero_bit(bitmap, bitmap_len,
1556 						   bit_a + 1) - 1;
1557 
1558 			if (!first)
1559 				proc_put_char(&buffer, &left, ',');
1560 			proc_put_long(&buffer, &left, bit_a, false);
1561 			if (bit_a != bit_b) {
1562 				proc_put_char(&buffer, &left, '-');
1563 				proc_put_long(&buffer, &left, bit_b, false);
1564 			}
1565 
1566 			first = 0; bit_b++;
1567 		}
1568 		proc_put_char(&buffer, &left, '\n');
1569 	}
1570 
1571 	if (!err) {
1572 		if (write) {
1573 			if (*ppos)
1574 				bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
1575 			else
1576 				bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
1577 		}
1578 		*lenp -= left;
1579 		*ppos += *lenp;
1580 	}
1581 
1582 	bitmap_free(tmp_bitmap);
1583 	return err;
1584 }
1585 
1586 #else /* CONFIG_PROC_SYSCTL */
1587 
proc_dostring(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1588 int proc_dostring(struct ctl_table *table, int write,
1589 		  void *buffer, size_t *lenp, loff_t *ppos)
1590 {
1591 	return -ENOSYS;
1592 }
1593 
proc_dointvec(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1594 int proc_dointvec(struct ctl_table *table, int write,
1595 		  void *buffer, size_t *lenp, loff_t *ppos)
1596 {
1597 	return -ENOSYS;
1598 }
1599 
proc_douintvec(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1600 int proc_douintvec(struct ctl_table *table, int write,
1601 		  void *buffer, size_t *lenp, loff_t *ppos)
1602 {
1603 	return -ENOSYS;
1604 }
1605 
proc_dointvec_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1606 int proc_dointvec_minmax(struct ctl_table *table, int write,
1607 		    void *buffer, size_t *lenp, loff_t *ppos)
1608 {
1609 	return -ENOSYS;
1610 }
1611 
proc_douintvec_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1612 int proc_douintvec_minmax(struct ctl_table *table, int write,
1613 			  void *buffer, size_t *lenp, loff_t *ppos)
1614 {
1615 	return -ENOSYS;
1616 }
1617 
proc_dointvec_jiffies(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1618 int proc_dointvec_jiffies(struct ctl_table *table, int write,
1619 		    void *buffer, size_t *lenp, loff_t *ppos)
1620 {
1621 	return -ENOSYS;
1622 }
1623 
proc_dointvec_userhz_jiffies(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1624 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1625 		    void *buffer, size_t *lenp, loff_t *ppos)
1626 {
1627 	return -ENOSYS;
1628 }
1629 
proc_dointvec_ms_jiffies(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1630 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
1631 			     void *buffer, size_t *lenp, loff_t *ppos)
1632 {
1633 	return -ENOSYS;
1634 }
1635 
proc_doulongvec_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1636 int proc_doulongvec_minmax(struct ctl_table *table, int write,
1637 		    void *buffer, size_t *lenp, loff_t *ppos)
1638 {
1639 	return -ENOSYS;
1640 }
1641 
proc_doulongvec_ms_jiffies_minmax(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1642 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1643 				      void *buffer, size_t *lenp, loff_t *ppos)
1644 {
1645 	return -ENOSYS;
1646 }
1647 
proc_do_large_bitmap(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1648 int proc_do_large_bitmap(struct ctl_table *table, int write,
1649 			 void *buffer, size_t *lenp, loff_t *ppos)
1650 {
1651 	return -ENOSYS;
1652 }
1653 
1654 #endif /* CONFIG_PROC_SYSCTL */
1655 
1656 #if defined(CONFIG_SYSCTL)
proc_do_static_key(struct ctl_table * table,int write,void * buffer,size_t * lenp,loff_t * ppos)1657 int proc_do_static_key(struct ctl_table *table, int write,
1658 		       void *buffer, size_t *lenp, loff_t *ppos)
1659 {
1660 	struct static_key *key = (struct static_key *)table->data;
1661 	static DEFINE_MUTEX(static_key_mutex);
1662 	int val, ret;
1663 	struct ctl_table tmp = {
1664 		.data   = &val,
1665 		.maxlen = sizeof(val),
1666 		.mode   = table->mode,
1667 		.extra1 = SYSCTL_ZERO,
1668 		.extra2 = SYSCTL_ONE,
1669 	};
1670 
1671 	if (write && !capable(CAP_SYS_ADMIN))
1672 		return -EPERM;
1673 
1674 	mutex_lock(&static_key_mutex);
1675 	val = static_key_enabled(key);
1676 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
1677 	if (write && !ret) {
1678 		if (val)
1679 			static_key_enable(key);
1680 		else
1681 			static_key_disable(key);
1682 	}
1683 	mutex_unlock(&static_key_mutex);
1684 	return ret;
1685 }
1686 
1687 static struct ctl_table kern_table[] = {
1688 	{
1689 		.procname	= "sched_child_runs_first",
1690 		.data		= &sysctl_sched_child_runs_first,
1691 		.maxlen		= sizeof(unsigned int),
1692 		.mode		= 0644,
1693 		.proc_handler	= proc_dointvec,
1694 	},
1695 #ifdef CONFIG_SCHED_DEBUG
1696 	{
1697 		.procname	= "sched_min_granularity_ns",
1698 		.data		= &sysctl_sched_min_granularity,
1699 		.maxlen		= sizeof(unsigned int),
1700 		.mode		= 0644,
1701 		.proc_handler	= sched_proc_update_handler,
1702 		.extra1		= &min_sched_granularity_ns,
1703 		.extra2		= &max_sched_granularity_ns,
1704 	},
1705 	{
1706 		.procname	= "sched_latency_ns",
1707 		.data		= &sysctl_sched_latency,
1708 		.maxlen		= sizeof(unsigned int),
1709 		.mode		= 0644,
1710 		.proc_handler	= sched_proc_update_handler,
1711 		.extra1		= &min_sched_granularity_ns,
1712 		.extra2		= &max_sched_granularity_ns,
1713 	},
1714 	{
1715 		.procname	= "sched_wakeup_granularity_ns",
1716 		.data		= &sysctl_sched_wakeup_granularity,
1717 		.maxlen		= sizeof(unsigned int),
1718 		.mode		= 0644,
1719 		.proc_handler	= sched_proc_update_handler,
1720 		.extra1		= &min_wakeup_granularity_ns,
1721 		.extra2		= &max_wakeup_granularity_ns,
1722 	},
1723 #ifdef CONFIG_SMP
1724 	{
1725 		.procname	= "sched_pelt_period",
1726 		.data		= &sysctl_sched_pelt_period,
1727 		.maxlen		= sizeof(int),
1728 		.mode		= 0644,
1729 		.proc_handler	= sched_pelt_period_update_handler,
1730 	},
1731 	{
1732 		.procname	= "sched_tunable_scaling",
1733 		.data		= &sysctl_sched_tunable_scaling,
1734 		.maxlen		= sizeof(enum sched_tunable_scaling),
1735 		.mode		= 0644,
1736 		.proc_handler	= sched_proc_update_handler,
1737 		.extra1		= &min_sched_tunable_scaling,
1738 		.extra2		= &max_sched_tunable_scaling,
1739 	},
1740 	{
1741 		.procname	= "sched_migration_cost_ns",
1742 		.data		= &sysctl_sched_migration_cost,
1743 		.maxlen		= sizeof(unsigned int),
1744 		.mode		= 0644,
1745 		.proc_handler	= proc_dointvec,
1746 	},
1747 	{
1748 		.procname	= "sched_nr_migrate",
1749 		.data		= &sysctl_sched_nr_migrate,
1750 		.maxlen		= sizeof(unsigned int),
1751 		.mode		= 0644,
1752 		.proc_handler	= proc_dointvec,
1753 	},
1754 #ifdef CONFIG_SCHEDSTATS
1755 	{
1756 		.procname	= "sched_schedstats",
1757 		.data		= NULL,
1758 		.maxlen		= sizeof(unsigned int),
1759 		.mode		= 0644,
1760 		.proc_handler	= sysctl_schedstats,
1761 		.extra1		= SYSCTL_ZERO,
1762 		.extra2		= SYSCTL_ONE,
1763 	},
1764 #endif /* CONFIG_SCHEDSTATS */
1765 #endif /* CONFIG_SMP */
1766 #ifdef CONFIG_NUMA_BALANCING
1767 	{
1768 		.procname	= "numa_balancing_scan_delay_ms",
1769 		.data		= &sysctl_numa_balancing_scan_delay,
1770 		.maxlen		= sizeof(unsigned int),
1771 		.mode		= 0644,
1772 		.proc_handler	= proc_dointvec,
1773 	},
1774 	{
1775 		.procname	= "numa_balancing_scan_period_min_ms",
1776 		.data		= &sysctl_numa_balancing_scan_period_min,
1777 		.maxlen		= sizeof(unsigned int),
1778 		.mode		= 0644,
1779 		.proc_handler	= proc_dointvec,
1780 	},
1781 	{
1782 		.procname	= "numa_balancing_scan_period_max_ms",
1783 		.data		= &sysctl_numa_balancing_scan_period_max,
1784 		.maxlen		= sizeof(unsigned int),
1785 		.mode		= 0644,
1786 		.proc_handler	= proc_dointvec,
1787 	},
1788 	{
1789 		.procname	= "numa_balancing_scan_size_mb",
1790 		.data		= &sysctl_numa_balancing_scan_size,
1791 		.maxlen		= sizeof(unsigned int),
1792 		.mode		= 0644,
1793 		.proc_handler	= proc_dointvec_minmax,
1794 		.extra1		= SYSCTL_ONE,
1795 	},
1796 	{
1797 		.procname	= "numa_balancing",
1798 		.data		= NULL, /* filled in by handler */
1799 		.maxlen		= sizeof(unsigned int),
1800 		.mode		= 0644,
1801 		.proc_handler	= sysctl_numa_balancing,
1802 		.extra1		= SYSCTL_ZERO,
1803 		.extra2		= SYSCTL_ONE,
1804 	},
1805 #endif /* CONFIG_NUMA_BALANCING */
1806 #endif /* CONFIG_SCHED_DEBUG */
1807 	{
1808 		.procname	= "sched_rt_period_us",
1809 		.data		= &sysctl_sched_rt_period,
1810 		.maxlen		= sizeof(unsigned int),
1811 		.mode		= 0644,
1812 		.proc_handler	= sched_rt_handler,
1813 	},
1814 	{
1815 		.procname	= "sched_rt_runtime_us",
1816 		.data		= &sysctl_sched_rt_runtime,
1817 		.maxlen		= sizeof(int),
1818 		.mode		= 0644,
1819 		.proc_handler	= sched_rt_handler,
1820 	},
1821 	{
1822 		.procname	= "sched_deadline_period_max_us",
1823 		.data		= &sysctl_sched_dl_period_max,
1824 		.maxlen		= sizeof(unsigned int),
1825 		.mode		= 0644,
1826 		.proc_handler	= proc_dointvec,
1827 	},
1828 	{
1829 		.procname	= "sched_deadline_period_min_us",
1830 		.data		= &sysctl_sched_dl_period_min,
1831 		.maxlen		= sizeof(unsigned int),
1832 		.mode		= 0644,
1833 		.proc_handler	= proc_dointvec,
1834 	},
1835 	{
1836 		.procname	= "sched_rr_timeslice_ms",
1837 		.data		= &sysctl_sched_rr_timeslice,
1838 		.maxlen		= sizeof(int),
1839 		.mode		= 0644,
1840 		.proc_handler	= sched_rr_handler,
1841 	},
1842 #ifdef CONFIG_SMP
1843 	{
1844 		.procname	= "sched_pelt_multiplier",
1845 		.data		= &sysctl_sched_pelt_multiplier,
1846 		.maxlen		= sizeof(unsigned int),
1847 		.mode		= 0644,
1848 		.proc_handler	= sched_pelt_multiplier,
1849 	},
1850 #endif
1851 #ifdef CONFIG_UCLAMP_TASK
1852 	{
1853 		.procname	= "sched_util_clamp_min",
1854 		.data		= &sysctl_sched_uclamp_util_min,
1855 		.maxlen		= sizeof(unsigned int),
1856 		.mode		= 0644,
1857 		.proc_handler	= sysctl_sched_uclamp_handler,
1858 	},
1859 	{
1860 		.procname	= "sched_util_clamp_max",
1861 		.data		= &sysctl_sched_uclamp_util_max,
1862 		.maxlen		= sizeof(unsigned int),
1863 		.mode		= 0644,
1864 		.proc_handler	= sysctl_sched_uclamp_handler,
1865 	},
1866 	{
1867 		.procname	= "sched_util_clamp_min_rt_default",
1868 		.data		= &sysctl_sched_uclamp_util_min_rt_default,
1869 		.maxlen		= sizeof(unsigned int),
1870 		.mode		= 0644,
1871 		.proc_handler	= sysctl_sched_uclamp_handler,
1872 	},
1873 #endif
1874 #ifdef CONFIG_SCHED_AUTOGROUP
1875 	{
1876 		.procname	= "sched_autogroup_enabled",
1877 		.data		= &sysctl_sched_autogroup_enabled,
1878 		.maxlen		= sizeof(unsigned int),
1879 		.mode		= 0644,
1880 		.proc_handler	= proc_dointvec_minmax,
1881 		.extra1		= SYSCTL_ZERO,
1882 		.extra2		= SYSCTL_ONE,
1883 	},
1884 #endif
1885 #ifdef CONFIG_CFS_BANDWIDTH
1886 	{
1887 		.procname	= "sched_cfs_bandwidth_slice_us",
1888 		.data		= &sysctl_sched_cfs_bandwidth_slice,
1889 		.maxlen		= sizeof(unsigned int),
1890 		.mode		= 0644,
1891 		.proc_handler	= proc_dointvec_minmax,
1892 		.extra1		= SYSCTL_ONE,
1893 	},
1894 #endif
1895 #if defined(CONFIG_ENERGY_MODEL) && defined(CONFIG_CPU_FREQ_GOV_SCHEDUTIL)
1896 	{
1897 		.procname	= "sched_energy_aware",
1898 		.data		= &sysctl_sched_energy_aware,
1899 		.maxlen		= sizeof(unsigned int),
1900 		.mode		= 0644,
1901 		.proc_handler	= sched_energy_aware_handler,
1902 		.extra1		= SYSCTL_ZERO,
1903 		.extra2		= SYSCTL_ONE,
1904 	},
1905 #endif
1906 #ifdef CONFIG_PROVE_LOCKING
1907 	{
1908 		.procname	= "prove_locking",
1909 		.data		= &prove_locking,
1910 		.maxlen		= sizeof(int),
1911 		.mode		= 0644,
1912 		.proc_handler	= proc_dointvec,
1913 	},
1914 #endif
1915 #ifdef CONFIG_LOCK_STAT
1916 	{
1917 		.procname	= "lock_stat",
1918 		.data		= &lock_stat,
1919 		.maxlen		= sizeof(int),
1920 		.mode		= 0644,
1921 		.proc_handler	= proc_dointvec,
1922 	},
1923 #endif
1924 	{
1925 		.procname	= "panic",
1926 		.data		= &panic_timeout,
1927 		.maxlen		= sizeof(int),
1928 		.mode		= 0644,
1929 		.proc_handler	= proc_dointvec,
1930 	},
1931 #ifdef CONFIG_COREDUMP
1932 	{
1933 		.procname	= "core_uses_pid",
1934 		.data		= &core_uses_pid,
1935 		.maxlen		= sizeof(int),
1936 		.mode		= 0644,
1937 		.proc_handler	= proc_dointvec,
1938 	},
1939 	{
1940 		.procname	= "core_pattern",
1941 		.data		= core_pattern,
1942 		.maxlen		= CORENAME_MAX_SIZE,
1943 		.mode		= 0644,
1944 		.proc_handler	= proc_dostring_coredump,
1945 	},
1946 	{
1947 		.procname	= "core_pipe_limit",
1948 		.data		= &core_pipe_limit,
1949 		.maxlen		= sizeof(unsigned int),
1950 		.mode		= 0644,
1951 		.proc_handler	= proc_dointvec,
1952 	},
1953 #endif
1954 #ifdef CONFIG_PROC_SYSCTL
1955 	{
1956 		.procname	= "tainted",
1957 		.maxlen 	= sizeof(long),
1958 		.mode		= 0644,
1959 		.proc_handler	= proc_taint,
1960 	},
1961 	{
1962 		.procname	= "sysctl_writes_strict",
1963 		.data		= &sysctl_writes_strict,
1964 		.maxlen		= sizeof(int),
1965 		.mode		= 0644,
1966 		.proc_handler	= proc_dointvec_minmax,
1967 		.extra1		= &neg_one,
1968 		.extra2		= SYSCTL_ONE,
1969 	},
1970 #endif
1971 #ifdef CONFIG_LATENCYTOP
1972 	{
1973 		.procname	= "latencytop",
1974 		.data		= &latencytop_enabled,
1975 		.maxlen		= sizeof(int),
1976 		.mode		= 0644,
1977 		.proc_handler	= sysctl_latencytop,
1978 	},
1979 #endif
1980 #ifdef CONFIG_BLK_DEV_INITRD
1981 	{
1982 		.procname	= "real-root-dev",
1983 		.data		= &real_root_dev,
1984 		.maxlen		= sizeof(int),
1985 		.mode		= 0644,
1986 		.proc_handler	= proc_dointvec,
1987 	},
1988 #endif
1989 	{
1990 		.procname	= "print-fatal-signals",
1991 		.data		= &print_fatal_signals,
1992 		.maxlen		= sizeof(int),
1993 		.mode		= 0644,
1994 		.proc_handler	= proc_dointvec,
1995 	},
1996 #ifdef CONFIG_SPARC
1997 	{
1998 		.procname	= "reboot-cmd",
1999 		.data		= reboot_command,
2000 		.maxlen		= 256,
2001 		.mode		= 0644,
2002 		.proc_handler	= proc_dostring,
2003 	},
2004 	{
2005 		.procname	= "stop-a",
2006 		.data		= &stop_a_enabled,
2007 		.maxlen		= sizeof (int),
2008 		.mode		= 0644,
2009 		.proc_handler	= proc_dointvec,
2010 	},
2011 	{
2012 		.procname	= "scons-poweroff",
2013 		.data		= &scons_pwroff,
2014 		.maxlen		= sizeof (int),
2015 		.mode		= 0644,
2016 		.proc_handler	= proc_dointvec,
2017 	},
2018 #endif
2019 #ifdef CONFIG_SPARC64
2020 	{
2021 		.procname	= "tsb-ratio",
2022 		.data		= &sysctl_tsb_ratio,
2023 		.maxlen		= sizeof (int),
2024 		.mode		= 0644,
2025 		.proc_handler	= proc_dointvec,
2026 	},
2027 #endif
2028 #ifdef CONFIG_PARISC
2029 	{
2030 		.procname	= "soft-power",
2031 		.data		= &pwrsw_enabled,
2032 		.maxlen		= sizeof (int),
2033 		.mode		= 0644,
2034 		.proc_handler	= proc_dointvec,
2035 	},
2036 #endif
2037 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
2038 	{
2039 		.procname	= "unaligned-trap",
2040 		.data		= &unaligned_enabled,
2041 		.maxlen		= sizeof (int),
2042 		.mode		= 0644,
2043 		.proc_handler	= proc_dointvec,
2044 	},
2045 #endif
2046 	{
2047 		.procname	= "ctrl-alt-del",
2048 		.data		= &C_A_D,
2049 		.maxlen		= sizeof(int),
2050 		.mode		= 0644,
2051 		.proc_handler	= proc_dointvec,
2052 	},
2053 #ifdef CONFIG_FUNCTION_TRACER
2054 	{
2055 		.procname	= "ftrace_enabled",
2056 		.data		= &ftrace_enabled,
2057 		.maxlen		= sizeof(int),
2058 		.mode		= 0644,
2059 		.proc_handler	= ftrace_enable_sysctl,
2060 	},
2061 #endif
2062 #ifdef CONFIG_STACK_TRACER
2063 	{
2064 		.procname	= "stack_tracer_enabled",
2065 		.data		= &stack_tracer_enabled,
2066 		.maxlen		= sizeof(int),
2067 		.mode		= 0644,
2068 		.proc_handler	= stack_trace_sysctl,
2069 	},
2070 #endif
2071 #ifdef CONFIG_TRACING
2072 	{
2073 		.procname	= "ftrace_dump_on_oops",
2074 		.data		= &ftrace_dump_on_oops,
2075 		.maxlen		= sizeof(int),
2076 		.mode		= 0644,
2077 		.proc_handler	= proc_dointvec,
2078 	},
2079 	{
2080 		.procname	= "traceoff_on_warning",
2081 		.data		= &__disable_trace_on_warning,
2082 		.maxlen		= sizeof(__disable_trace_on_warning),
2083 		.mode		= 0644,
2084 		.proc_handler	= proc_dointvec,
2085 	},
2086 	{
2087 		.procname	= "tracepoint_printk",
2088 		.data		= &tracepoint_printk,
2089 		.maxlen		= sizeof(tracepoint_printk),
2090 		.mode		= 0644,
2091 		.proc_handler	= tracepoint_printk_sysctl,
2092 	},
2093 #endif
2094 #ifdef CONFIG_KEXEC_CORE
2095 	{
2096 		.procname	= "kexec_load_disabled",
2097 		.data		= &kexec_load_disabled,
2098 		.maxlen		= sizeof(int),
2099 		.mode		= 0644,
2100 		/* only handle a transition from default "0" to "1" */
2101 		.proc_handler	= proc_dointvec_minmax,
2102 		.extra1		= SYSCTL_ONE,
2103 		.extra2		= SYSCTL_ONE,
2104 	},
2105 #endif
2106 #ifdef CONFIG_MODULES
2107 	{
2108 		.procname	= "modprobe",
2109 		.data		= &modprobe_path,
2110 		.maxlen		= KMOD_PATH_LEN,
2111 		.mode		= 0644,
2112 		.proc_handler	= proc_dostring,
2113 	},
2114 	{
2115 		.procname	= "modules_disabled",
2116 		.data		= &modules_disabled,
2117 		.maxlen		= sizeof(int),
2118 		.mode		= 0644,
2119 		/* only handle a transition from default "0" to "1" */
2120 		.proc_handler	= proc_dointvec_minmax,
2121 		.extra1		= SYSCTL_ONE,
2122 		.extra2		= SYSCTL_ONE,
2123 	},
2124 #endif
2125 #ifdef CONFIG_UEVENT_HELPER
2126 	{
2127 		.procname	= "hotplug",
2128 		.data		= &uevent_helper,
2129 		.maxlen		= UEVENT_HELPER_PATH_LEN,
2130 		.mode		= 0644,
2131 		.proc_handler	= proc_dostring,
2132 	},
2133 #endif
2134 #ifdef CONFIG_CHR_DEV_SG
2135 	{
2136 		.procname	= "sg-big-buff",
2137 		.data		= &sg_big_buff,
2138 		.maxlen		= sizeof (int),
2139 		.mode		= 0444,
2140 		.proc_handler	= proc_dointvec,
2141 	},
2142 #endif
2143 #ifdef CONFIG_BSD_PROCESS_ACCT
2144 	{
2145 		.procname	= "acct",
2146 		.data		= &acct_parm,
2147 		.maxlen		= 3*sizeof(int),
2148 		.mode		= 0644,
2149 		.proc_handler	= proc_dointvec,
2150 	},
2151 #endif
2152 #ifdef CONFIG_MAGIC_SYSRQ
2153 	{
2154 		.procname	= "sysrq",
2155 		.data		= NULL,
2156 		.maxlen		= sizeof (int),
2157 		.mode		= 0644,
2158 		.proc_handler	= sysrq_sysctl_handler,
2159 	},
2160 #endif
2161 #ifdef CONFIG_PROC_SYSCTL
2162 	{
2163 		.procname	= "cad_pid",
2164 		.data		= NULL,
2165 		.maxlen		= sizeof (int),
2166 		.mode		= 0600,
2167 		.proc_handler	= proc_do_cad_pid,
2168 	},
2169 #endif
2170 	{
2171 		.procname	= "threads-max",
2172 		.data		= NULL,
2173 		.maxlen		= sizeof(int),
2174 		.mode		= 0644,
2175 		.proc_handler	= sysctl_max_threads,
2176 	},
2177 	{
2178 		.procname	= "random",
2179 		.mode		= 0555,
2180 		.child		= random_table,
2181 	},
2182 	{
2183 		.procname	= "usermodehelper",
2184 		.mode		= 0555,
2185 		.child		= usermodehelper_table,
2186 	},
2187 #ifdef CONFIG_FW_LOADER_USER_HELPER
2188 	{
2189 		.procname	= "firmware_config",
2190 		.mode		= 0555,
2191 		.child		= firmware_config_table,
2192 	},
2193 #endif
2194 	{
2195 		.procname	= "overflowuid",
2196 		.data		= &overflowuid,
2197 		.maxlen		= sizeof(int),
2198 		.mode		= 0644,
2199 		.proc_handler	= proc_dointvec_minmax,
2200 		.extra1		= &minolduid,
2201 		.extra2		= &maxolduid,
2202 	},
2203 	{
2204 		.procname	= "overflowgid",
2205 		.data		= &overflowgid,
2206 		.maxlen		= sizeof(int),
2207 		.mode		= 0644,
2208 		.proc_handler	= proc_dointvec_minmax,
2209 		.extra1		= &minolduid,
2210 		.extra2		= &maxolduid,
2211 	},
2212 #ifdef CONFIG_S390
2213 	{
2214 		.procname	= "userprocess_debug",
2215 		.data		= &show_unhandled_signals,
2216 		.maxlen		= sizeof(int),
2217 		.mode		= 0644,
2218 		.proc_handler	= proc_dointvec,
2219 	},
2220 #endif
2221 #ifdef CONFIG_SMP
2222 	{
2223 		.procname	= "oops_all_cpu_backtrace",
2224 		.data		= &sysctl_oops_all_cpu_backtrace,
2225 		.maxlen		= sizeof(int),
2226 		.mode		= 0644,
2227 		.proc_handler	= proc_dointvec_minmax,
2228 		.extra1		= SYSCTL_ZERO,
2229 		.extra2		= SYSCTL_ONE,
2230 	},
2231 #endif /* CONFIG_SMP */
2232 	{
2233 		.procname	= "pid_max",
2234 		.data		= &pid_max,
2235 		.maxlen		= sizeof (int),
2236 		.mode		= 0644,
2237 		.proc_handler	= proc_dointvec_minmax,
2238 		.extra1		= &pid_max_min,
2239 		.extra2		= &pid_max_max,
2240 	},
2241 	{
2242 		.procname	= "panic_on_oops",
2243 		.data		= &panic_on_oops,
2244 		.maxlen		= sizeof(int),
2245 		.mode		= 0644,
2246 		.proc_handler	= proc_dointvec,
2247 	},
2248 	{
2249 		.procname	= "panic_print",
2250 		.data		= &panic_print,
2251 		.maxlen		= sizeof(unsigned long),
2252 		.mode		= 0644,
2253 		.proc_handler	= proc_doulongvec_minmax,
2254 	},
2255 #if defined CONFIG_PRINTK
2256 	{
2257 		.procname	= "printk",
2258 		.data		= &console_loglevel,
2259 		.maxlen		= 4*sizeof(int),
2260 		.mode		= 0644,
2261 		.proc_handler	= proc_dointvec,
2262 	},
2263 	{
2264 		.procname	= "printk_ratelimit",
2265 		.data		= &printk_ratelimit_state.interval,
2266 		.maxlen		= sizeof(int),
2267 		.mode		= 0644,
2268 		.proc_handler	= proc_dointvec_jiffies,
2269 	},
2270 	{
2271 		.procname	= "printk_ratelimit_burst",
2272 		.data		= &printk_ratelimit_state.burst,
2273 		.maxlen		= sizeof(int),
2274 		.mode		= 0644,
2275 		.proc_handler	= proc_dointvec,
2276 	},
2277 	{
2278 		.procname	= "printk_delay",
2279 		.data		= &printk_delay_msec,
2280 		.maxlen		= sizeof(int),
2281 		.mode		= 0644,
2282 		.proc_handler	= proc_dointvec_minmax,
2283 		.extra1		= SYSCTL_ZERO,
2284 		.extra2		= &ten_thousand,
2285 	},
2286 	{
2287 		.procname	= "printk_devkmsg",
2288 		.data		= devkmsg_log_str,
2289 		.maxlen		= DEVKMSG_STR_MAX_SIZE,
2290 		.mode		= 0644,
2291 		.proc_handler	= devkmsg_sysctl_set_loglvl,
2292 	},
2293 	{
2294 		.procname	= "dmesg_restrict",
2295 		.data		= &dmesg_restrict,
2296 		.maxlen		= sizeof(int),
2297 		.mode		= 0644,
2298 		.proc_handler	= proc_dointvec_minmax_sysadmin,
2299 		.extra1		= SYSCTL_ZERO,
2300 		.extra2		= SYSCTL_ONE,
2301 	},
2302 	{
2303 		.procname	= "kptr_restrict",
2304 		.data		= &kptr_restrict,
2305 		.maxlen		= sizeof(int),
2306 		.mode		= 0644,
2307 		.proc_handler	= proc_dointvec_minmax_sysadmin,
2308 		.extra1		= SYSCTL_ZERO,
2309 		.extra2		= &two,
2310 	},
2311 #endif
2312 	{
2313 		.procname	= "ngroups_max",
2314 		.data		= &ngroups_max,
2315 		.maxlen		= sizeof (int),
2316 		.mode		= 0444,
2317 		.proc_handler	= proc_dointvec,
2318 	},
2319 	{
2320 		.procname	= "cap_last_cap",
2321 		.data		= (void *)&cap_last_cap,
2322 		.maxlen		= sizeof(int),
2323 		.mode		= 0444,
2324 		.proc_handler	= proc_dointvec,
2325 	},
2326 #if defined(CONFIG_LOCKUP_DETECTOR)
2327 	{
2328 		.procname       = "watchdog",
2329 		.data		= &watchdog_user_enabled,
2330 		.maxlen		= sizeof(int),
2331 		.mode		= 0644,
2332 		.proc_handler   = proc_watchdog,
2333 		.extra1		= SYSCTL_ZERO,
2334 		.extra2		= SYSCTL_ONE,
2335 	},
2336 	{
2337 		.procname	= "watchdog_thresh",
2338 		.data		= &watchdog_thresh,
2339 		.maxlen		= sizeof(int),
2340 		.mode		= 0644,
2341 		.proc_handler	= proc_watchdog_thresh,
2342 		.extra1		= SYSCTL_ZERO,
2343 		.extra2		= &sixty,
2344 	},
2345 	{
2346 		.procname       = "nmi_watchdog",
2347 		.data		= &nmi_watchdog_user_enabled,
2348 		.maxlen		= sizeof(int),
2349 		.mode		= NMI_WATCHDOG_SYSCTL_PERM,
2350 		.proc_handler   = proc_nmi_watchdog,
2351 		.extra1		= SYSCTL_ZERO,
2352 		.extra2		= SYSCTL_ONE,
2353 	},
2354 	{
2355 		.procname	= "watchdog_cpumask",
2356 		.data		= &watchdog_cpumask_bits,
2357 		.maxlen		= NR_CPUS,
2358 		.mode		= 0644,
2359 		.proc_handler	= proc_watchdog_cpumask,
2360 	},
2361 #ifdef CONFIG_SOFTLOCKUP_DETECTOR
2362 	{
2363 		.procname       = "soft_watchdog",
2364 		.data		= &soft_watchdog_user_enabled,
2365 		.maxlen		= sizeof(int),
2366 		.mode		= 0644,
2367 		.proc_handler   = proc_soft_watchdog,
2368 		.extra1		= SYSCTL_ZERO,
2369 		.extra2		= SYSCTL_ONE,
2370 	},
2371 	{
2372 		.procname	= "softlockup_panic",
2373 		.data		= &softlockup_panic,
2374 		.maxlen		= sizeof(int),
2375 		.mode		= 0644,
2376 		.proc_handler	= proc_dointvec_minmax,
2377 		.extra1		= SYSCTL_ZERO,
2378 		.extra2		= SYSCTL_ONE,
2379 	},
2380 #ifdef CONFIG_SMP
2381 	{
2382 		.procname	= "softlockup_all_cpu_backtrace",
2383 		.data		= &sysctl_softlockup_all_cpu_backtrace,
2384 		.maxlen		= sizeof(int),
2385 		.mode		= 0644,
2386 		.proc_handler	= proc_dointvec_minmax,
2387 		.extra1		= SYSCTL_ZERO,
2388 		.extra2		= SYSCTL_ONE,
2389 	},
2390 #endif /* CONFIG_SMP */
2391 #endif
2392 #ifdef CONFIG_HARDLOCKUP_DETECTOR
2393 	{
2394 		.procname	= "hardlockup_panic",
2395 		.data		= &hardlockup_panic,
2396 		.maxlen		= sizeof(int),
2397 		.mode		= 0644,
2398 		.proc_handler	= proc_dointvec_minmax,
2399 		.extra1		= SYSCTL_ZERO,
2400 		.extra2		= SYSCTL_ONE,
2401 	},
2402 #ifdef CONFIG_SMP
2403 	{
2404 		.procname	= "hardlockup_all_cpu_backtrace",
2405 		.data		= &sysctl_hardlockup_all_cpu_backtrace,
2406 		.maxlen		= sizeof(int),
2407 		.mode		= 0644,
2408 		.proc_handler	= proc_dointvec_minmax,
2409 		.extra1		= SYSCTL_ZERO,
2410 		.extra2		= SYSCTL_ONE,
2411 	},
2412 #endif /* CONFIG_SMP */
2413 #endif
2414 #endif
2415 
2416 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
2417 	{
2418 		.procname       = "unknown_nmi_panic",
2419 		.data           = &unknown_nmi_panic,
2420 		.maxlen         = sizeof (int),
2421 		.mode           = 0644,
2422 		.proc_handler   = proc_dointvec,
2423 	},
2424 #endif
2425 
2426 #if (defined(CONFIG_X86_32) || defined(CONFIG_PARISC)) && \
2427 	defined(CONFIG_DEBUG_STACKOVERFLOW)
2428 	{
2429 		.procname	= "panic_on_stackoverflow",
2430 		.data		= &sysctl_panic_on_stackoverflow,
2431 		.maxlen		= sizeof(int),
2432 		.mode		= 0644,
2433 		.proc_handler	= proc_dointvec,
2434 	},
2435 #endif
2436 #if defined(CONFIG_X86)
2437 	{
2438 		.procname	= "panic_on_unrecovered_nmi",
2439 		.data		= &panic_on_unrecovered_nmi,
2440 		.maxlen		= sizeof(int),
2441 		.mode		= 0644,
2442 		.proc_handler	= proc_dointvec,
2443 	},
2444 	{
2445 		.procname	= "panic_on_io_nmi",
2446 		.data		= &panic_on_io_nmi,
2447 		.maxlen		= sizeof(int),
2448 		.mode		= 0644,
2449 		.proc_handler	= proc_dointvec,
2450 	},
2451 	{
2452 		.procname	= "bootloader_type",
2453 		.data		= &bootloader_type,
2454 		.maxlen		= sizeof (int),
2455 		.mode		= 0444,
2456 		.proc_handler	= proc_dointvec,
2457 	},
2458 	{
2459 		.procname	= "bootloader_version",
2460 		.data		= &bootloader_version,
2461 		.maxlen		= sizeof (int),
2462 		.mode		= 0444,
2463 		.proc_handler	= proc_dointvec,
2464 	},
2465 	{
2466 		.procname	= "io_delay_type",
2467 		.data		= &io_delay_type,
2468 		.maxlen		= sizeof(int),
2469 		.mode		= 0644,
2470 		.proc_handler	= proc_dointvec,
2471 	},
2472 #endif
2473 #if defined(CONFIG_MMU)
2474 	{
2475 		.procname	= "randomize_va_space",
2476 		.data		= &randomize_va_space,
2477 		.maxlen		= sizeof(int),
2478 		.mode		= 0644,
2479 		.proc_handler	= proc_dointvec,
2480 	},
2481 #endif
2482 #if defined(CONFIG_S390) && defined(CONFIG_SMP)
2483 	{
2484 		.procname	= "spin_retry",
2485 		.data		= &spin_retry,
2486 		.maxlen		= sizeof (int),
2487 		.mode		= 0644,
2488 		.proc_handler	= proc_dointvec,
2489 	},
2490 #endif
2491 #if	defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
2492 	{
2493 		.procname	= "acpi_video_flags",
2494 		.data		= &acpi_realmode_flags,
2495 		.maxlen		= sizeof (unsigned long),
2496 		.mode		= 0644,
2497 		.proc_handler	= proc_doulongvec_minmax,
2498 	},
2499 #endif
2500 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
2501 	{
2502 		.procname	= "ignore-unaligned-usertrap",
2503 		.data		= &no_unaligned_warning,
2504 		.maxlen		= sizeof (int),
2505 		.mode		= 0644,
2506 		.proc_handler	= proc_dointvec,
2507 	},
2508 #endif
2509 #ifdef CONFIG_IA64
2510 	{
2511 		.procname	= "unaligned-dump-stack",
2512 		.data		= &unaligned_dump_stack,
2513 		.maxlen		= sizeof (int),
2514 		.mode		= 0644,
2515 		.proc_handler	= proc_dointvec,
2516 	},
2517 #endif
2518 #ifdef CONFIG_DETECT_HUNG_TASK
2519 #ifdef CONFIG_SMP
2520 	{
2521 		.procname	= "hung_task_all_cpu_backtrace",
2522 		.data		= &sysctl_hung_task_all_cpu_backtrace,
2523 		.maxlen		= sizeof(int),
2524 		.mode		= 0644,
2525 		.proc_handler	= proc_dointvec_minmax,
2526 		.extra1		= SYSCTL_ZERO,
2527 		.extra2		= SYSCTL_ONE,
2528 	},
2529 #endif /* CONFIG_SMP */
2530 	{
2531 		.procname	= "hung_task_panic",
2532 		.data		= &sysctl_hung_task_panic,
2533 		.maxlen		= sizeof(int),
2534 		.mode		= 0644,
2535 		.proc_handler	= proc_dointvec_minmax,
2536 		.extra1		= SYSCTL_ZERO,
2537 		.extra2		= SYSCTL_ONE,
2538 	},
2539 	{
2540 		.procname	= "hung_task_check_count",
2541 		.data		= &sysctl_hung_task_check_count,
2542 		.maxlen		= sizeof(int),
2543 		.mode		= 0644,
2544 		.proc_handler	= proc_dointvec_minmax,
2545 		.extra1		= SYSCTL_ZERO,
2546 	},
2547 	{
2548 		.procname	= "hung_task_timeout_secs",
2549 		.data		= &sysctl_hung_task_timeout_secs,
2550 		.maxlen		= sizeof(unsigned long),
2551 		.mode		= 0644,
2552 		.proc_handler	= proc_dohung_task_timeout_secs,
2553 		.extra2		= &hung_task_timeout_max,
2554 	},
2555 	{
2556 		.procname	= "hung_task_check_interval_secs",
2557 		.data		= &sysctl_hung_task_check_interval_secs,
2558 		.maxlen		= sizeof(unsigned long),
2559 		.mode		= 0644,
2560 		.proc_handler	= proc_dohung_task_timeout_secs,
2561 		.extra2		= &hung_task_timeout_max,
2562 	},
2563 	{
2564 		.procname	= "hung_task_warnings",
2565 		.data		= &sysctl_hung_task_warnings,
2566 		.maxlen		= sizeof(int),
2567 		.mode		= 0644,
2568 		.proc_handler	= proc_dointvec_minmax,
2569 		.extra1		= &neg_one,
2570 	},
2571 #endif
2572 #ifdef CONFIG_RT_MUTEXES
2573 	{
2574 		.procname	= "max_lock_depth",
2575 		.data		= &max_lock_depth,
2576 		.maxlen		= sizeof(int),
2577 		.mode		= 0644,
2578 		.proc_handler	= proc_dointvec,
2579 	},
2580 #endif
2581 	{
2582 		.procname	= "poweroff_cmd",
2583 		.data		= &poweroff_cmd,
2584 		.maxlen		= POWEROFF_CMD_PATH_LEN,
2585 		.mode		= 0644,
2586 		.proc_handler	= proc_dostring,
2587 	},
2588 #ifdef CONFIG_KEYS
2589 	{
2590 		.procname	= "keys",
2591 		.mode		= 0555,
2592 		.child		= key_sysctls,
2593 	},
2594 #endif
2595 #ifdef CONFIG_PERF_EVENTS
2596 	/*
2597 	 * User-space scripts rely on the existence of this file
2598 	 * as a feature check for perf_events being enabled.
2599 	 *
2600 	 * So it's an ABI, do not remove!
2601 	 */
2602 	{
2603 		.procname	= "perf_event_paranoid",
2604 		.data		= &sysctl_perf_event_paranoid,
2605 		.maxlen		= sizeof(sysctl_perf_event_paranoid),
2606 		.mode		= 0644,
2607 		.proc_handler	= proc_dointvec,
2608 	},
2609 	{
2610 		.procname	= "perf_event_mlock_kb",
2611 		.data		= &sysctl_perf_event_mlock,
2612 		.maxlen		= sizeof(sysctl_perf_event_mlock),
2613 		.mode		= 0644,
2614 		.proc_handler	= proc_dointvec,
2615 	},
2616 	{
2617 		.procname	= "perf_event_max_sample_rate",
2618 		.data		= &sysctl_perf_event_sample_rate,
2619 		.maxlen		= sizeof(sysctl_perf_event_sample_rate),
2620 		.mode		= 0644,
2621 		.proc_handler	= perf_proc_update_handler,
2622 		.extra1		= SYSCTL_ONE,
2623 	},
2624 	{
2625 		.procname	= "perf_cpu_time_max_percent",
2626 		.data		= &sysctl_perf_cpu_time_max_percent,
2627 		.maxlen		= sizeof(sysctl_perf_cpu_time_max_percent),
2628 		.mode		= 0644,
2629 		.proc_handler	= perf_cpu_time_max_percent_handler,
2630 		.extra1		= SYSCTL_ZERO,
2631 		.extra2		= &one_hundred,
2632 	},
2633 	{
2634 		.procname	= "perf_event_max_stack",
2635 		.data		= &sysctl_perf_event_max_stack,
2636 		.maxlen		= sizeof(sysctl_perf_event_max_stack),
2637 		.mode		= 0644,
2638 		.proc_handler	= perf_event_max_stack_handler,
2639 		.extra1		= SYSCTL_ZERO,
2640 		.extra2		= &six_hundred_forty_kb,
2641 	},
2642 	{
2643 		.procname	= "perf_event_max_contexts_per_stack",
2644 		.data		= &sysctl_perf_event_max_contexts_per_stack,
2645 		.maxlen		= sizeof(sysctl_perf_event_max_contexts_per_stack),
2646 		.mode		= 0644,
2647 		.proc_handler	= perf_event_max_stack_handler,
2648 		.extra1		= SYSCTL_ZERO,
2649 		.extra2		= &one_thousand,
2650 	},
2651 #endif
2652 	{
2653 		.procname	= "panic_on_warn",
2654 		.data		= &panic_on_warn,
2655 		.maxlen		= sizeof(int),
2656 		.mode		= 0644,
2657 		.proc_handler	= proc_dointvec_minmax,
2658 		.extra1		= SYSCTL_ZERO,
2659 		.extra2		= SYSCTL_ONE,
2660 	},
2661 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
2662 	{
2663 		.procname	= "timer_migration",
2664 		.data		= &sysctl_timer_migration,
2665 		.maxlen		= sizeof(unsigned int),
2666 		.mode		= 0644,
2667 		.proc_handler	= timer_migration_handler,
2668 		.extra1		= SYSCTL_ZERO,
2669 		.extra2		= SYSCTL_ONE,
2670 	},
2671 #endif
2672 #ifdef CONFIG_BPF_SYSCALL
2673 	{
2674 		.procname	= "unprivileged_bpf_disabled",
2675 		.data		= &sysctl_unprivileged_bpf_disabled,
2676 		.maxlen		= sizeof(sysctl_unprivileged_bpf_disabled),
2677 		.mode		= 0644,
2678 		.proc_handler	= bpf_unpriv_handler,
2679 		.extra1		= SYSCTL_ZERO,
2680 		.extra2		= &two,
2681 	},
2682 	{
2683 		.procname	= "bpf_stats_enabled",
2684 		.data		= &bpf_stats_enabled_key.key,
2685 		.maxlen		= sizeof(bpf_stats_enabled_key),
2686 		.mode		= 0644,
2687 		.proc_handler	= bpf_stats_handler,
2688 	},
2689 #endif
2690 #if defined(CONFIG_TREE_RCU)
2691 	{
2692 		.procname	= "panic_on_rcu_stall",
2693 		.data		= &sysctl_panic_on_rcu_stall,
2694 		.maxlen		= sizeof(sysctl_panic_on_rcu_stall),
2695 		.mode		= 0644,
2696 		.proc_handler	= proc_dointvec_minmax,
2697 		.extra1		= SYSCTL_ZERO,
2698 		.extra2		= SYSCTL_ONE,
2699 	},
2700 #endif
2701 #ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
2702 	{
2703 		.procname	= "stack_erasing",
2704 		.data		= NULL,
2705 		.maxlen		= sizeof(int),
2706 		.mode		= 0600,
2707 		.proc_handler	= stack_erasing_sysctl,
2708 		.extra1		= SYSCTL_ZERO,
2709 		.extra2		= SYSCTL_ONE,
2710 	},
2711 #endif
2712 	{ }
2713 };
2714 
2715 static struct ctl_table vm_table[] = {
2716 	{
2717 		.procname	= "overcommit_memory",
2718 		.data		= &sysctl_overcommit_memory,
2719 		.maxlen		= sizeof(sysctl_overcommit_memory),
2720 		.mode		= 0644,
2721 		.proc_handler	= overcommit_policy_handler,
2722 		.extra1		= SYSCTL_ZERO,
2723 		.extra2		= &two,
2724 	},
2725 	{
2726 		.procname	= "panic_on_oom",
2727 		.data		= &sysctl_panic_on_oom,
2728 		.maxlen		= sizeof(sysctl_panic_on_oom),
2729 		.mode		= 0644,
2730 		.proc_handler	= proc_dointvec_minmax,
2731 		.extra1		= SYSCTL_ZERO,
2732 		.extra2		= &two,
2733 	},
2734 	{
2735 		.procname	= "oom_kill_allocating_task",
2736 		.data		= &sysctl_oom_kill_allocating_task,
2737 		.maxlen		= sizeof(sysctl_oom_kill_allocating_task),
2738 		.mode		= 0644,
2739 		.proc_handler	= proc_dointvec,
2740 	},
2741 	{
2742 		.procname	= "oom_dump_tasks",
2743 		.data		= &sysctl_oom_dump_tasks,
2744 		.maxlen		= sizeof(sysctl_oom_dump_tasks),
2745 		.mode		= 0644,
2746 		.proc_handler	= proc_dointvec,
2747 	},
2748 	{
2749 		.procname	= "overcommit_ratio",
2750 		.data		= &sysctl_overcommit_ratio,
2751 		.maxlen		= sizeof(sysctl_overcommit_ratio),
2752 		.mode		= 0644,
2753 		.proc_handler	= overcommit_ratio_handler,
2754 	},
2755 	{
2756 		.procname	= "overcommit_kbytes",
2757 		.data		= &sysctl_overcommit_kbytes,
2758 		.maxlen		= sizeof(sysctl_overcommit_kbytes),
2759 		.mode		= 0644,
2760 		.proc_handler	= overcommit_kbytes_handler,
2761 	},
2762 	{
2763 		.procname	= "page-cluster",
2764 		.data		= &page_cluster,
2765 		.maxlen		= sizeof(int),
2766 		.mode		= 0644,
2767 		.proc_handler	= proc_dointvec_minmax,
2768 		.extra1		= SYSCTL_ZERO,
2769 	},
2770 	{
2771 		.procname	= "dirty_background_ratio",
2772 		.data		= &dirty_background_ratio,
2773 		.maxlen		= sizeof(dirty_background_ratio),
2774 		.mode		= 0644,
2775 		.proc_handler	= dirty_background_ratio_handler,
2776 		.extra1		= SYSCTL_ZERO,
2777 		.extra2		= &one_hundred,
2778 	},
2779 	{
2780 		.procname	= "dirty_background_bytes",
2781 		.data		= &dirty_background_bytes,
2782 		.maxlen		= sizeof(dirty_background_bytes),
2783 		.mode		= 0644,
2784 		.proc_handler	= dirty_background_bytes_handler,
2785 		.extra1		= &one_ul,
2786 	},
2787 	{
2788 		.procname	= "dirty_ratio",
2789 		.data		= &vm_dirty_ratio,
2790 		.maxlen		= sizeof(vm_dirty_ratio),
2791 		.mode		= 0644,
2792 		.proc_handler	= dirty_ratio_handler,
2793 		.extra1		= SYSCTL_ZERO,
2794 		.extra2		= &one_hundred,
2795 	},
2796 	{
2797 		.procname	= "dirty_bytes",
2798 		.data		= &vm_dirty_bytes,
2799 		.maxlen		= sizeof(vm_dirty_bytes),
2800 		.mode		= 0644,
2801 		.proc_handler	= dirty_bytes_handler,
2802 		.extra1		= &dirty_bytes_min,
2803 	},
2804 	{
2805 		.procname	= "dirty_writeback_centisecs",
2806 		.data		= &dirty_writeback_interval,
2807 		.maxlen		= sizeof(dirty_writeback_interval),
2808 		.mode		= 0644,
2809 		.proc_handler	= dirty_writeback_centisecs_handler,
2810 	},
2811 	{
2812 		.procname	= "dirty_expire_centisecs",
2813 		.data		= &dirty_expire_interval,
2814 		.maxlen		= sizeof(dirty_expire_interval),
2815 		.mode		= 0644,
2816 		.proc_handler	= proc_dointvec_minmax,
2817 		.extra1		= SYSCTL_ZERO,
2818 	},
2819 	{
2820 		.procname	= "dirtytime_expire_seconds",
2821 		.data		= &dirtytime_expire_interval,
2822 		.maxlen		= sizeof(dirtytime_expire_interval),
2823 		.mode		= 0644,
2824 		.proc_handler	= dirtytime_interval_handler,
2825 		.extra1		= SYSCTL_ZERO,
2826 	},
2827 	{
2828 		.procname	= "swappiness",
2829 		.data		= &vm_swappiness,
2830 		.maxlen		= sizeof(vm_swappiness),
2831 		.mode		= 0644,
2832 		.proc_handler	= proc_dointvec_minmax,
2833 		.extra1		= SYSCTL_ZERO,
2834 		.extra2		= &two_hundred,
2835 	},
2836 #ifdef CONFIG_NUMA
2837 	{
2838 		.procname	= "numa_stat",
2839 		.data		= &sysctl_vm_numa_stat,
2840 		.maxlen		= sizeof(int),
2841 		.mode		= 0644,
2842 		.proc_handler	= sysctl_vm_numa_stat_handler,
2843 		.extra1		= SYSCTL_ZERO,
2844 		.extra2		= SYSCTL_ONE,
2845 	},
2846 #endif
2847 #ifdef CONFIG_HUGETLB_PAGE
2848 	{
2849 		.procname	= "nr_hugepages",
2850 		.data		= NULL,
2851 		.maxlen		= sizeof(unsigned long),
2852 		.mode		= 0644,
2853 		.proc_handler	= hugetlb_sysctl_handler,
2854 	},
2855 #ifdef CONFIG_NUMA
2856 	{
2857 		.procname       = "nr_hugepages_mempolicy",
2858 		.data           = NULL,
2859 		.maxlen         = sizeof(unsigned long),
2860 		.mode           = 0644,
2861 		.proc_handler   = &hugetlb_mempolicy_sysctl_handler,
2862 	},
2863 #endif
2864 	 {
2865 		.procname	= "hugetlb_shm_group",
2866 		.data		= &sysctl_hugetlb_shm_group,
2867 		.maxlen		= sizeof(gid_t),
2868 		.mode		= 0644,
2869 		.proc_handler	= proc_dointvec,
2870 	 },
2871 	{
2872 		.procname	= "nr_overcommit_hugepages",
2873 		.data		= NULL,
2874 		.maxlen		= sizeof(unsigned long),
2875 		.mode		= 0644,
2876 		.proc_handler	= hugetlb_overcommit_handler,
2877 	},
2878 #endif
2879 	{
2880 		.procname	= "lowmem_reserve_ratio",
2881 		.data		= &sysctl_lowmem_reserve_ratio,
2882 		.maxlen		= sizeof(sysctl_lowmem_reserve_ratio),
2883 		.mode		= 0644,
2884 		.proc_handler	= lowmem_reserve_ratio_sysctl_handler,
2885 	},
2886 	{
2887 		.procname	= "drop_caches",
2888 		.data		= &sysctl_drop_caches,
2889 		.maxlen		= sizeof(int),
2890 		.mode		= 0200,
2891 		.proc_handler	= drop_caches_sysctl_handler,
2892 		.extra1		= SYSCTL_ONE,
2893 		.extra2		= &four,
2894 	},
2895 #ifdef CONFIG_COMPACTION
2896 	{
2897 		.procname	= "compact_memory",
2898 		.data		= &sysctl_compact_memory,
2899 		.maxlen		= sizeof(int),
2900 		.mode		= 0200,
2901 		.proc_handler	= sysctl_compaction_handler,
2902 	},
2903 	{
2904 		.procname	= "compaction_proactiveness",
2905 		.data		= &sysctl_compaction_proactiveness,
2906 		.maxlen		= sizeof(sysctl_compaction_proactiveness),
2907 		.mode		= 0644,
2908 		.proc_handler	= compaction_proactiveness_sysctl_handler,
2909 		.extra1		= SYSCTL_ZERO,
2910 		.extra2		= &one_hundred,
2911 	},
2912 	{
2913 		.procname	= "extfrag_threshold",
2914 		.data		= &sysctl_extfrag_threshold,
2915 		.maxlen		= sizeof(int),
2916 		.mode		= 0644,
2917 		.proc_handler	= proc_dointvec_minmax,
2918 		.extra1		= &min_extfrag_threshold,
2919 		.extra2		= &max_extfrag_threshold,
2920 	},
2921 	{
2922 		.procname	= "compact_unevictable_allowed",
2923 		.data		= &sysctl_compact_unevictable_allowed,
2924 		.maxlen		= sizeof(int),
2925 		.mode		= 0644,
2926 		.proc_handler	= proc_dointvec_minmax_warn_RT_change,
2927 		.extra1		= SYSCTL_ZERO,
2928 		.extra2		= SYSCTL_ONE,
2929 	},
2930 
2931 #endif /* CONFIG_COMPACTION */
2932 	{
2933 		.procname	= "min_free_kbytes",
2934 		.data		= &min_free_kbytes,
2935 		.maxlen		= sizeof(min_free_kbytes),
2936 		.mode		= 0644,
2937 		.proc_handler	= min_free_kbytes_sysctl_handler,
2938 		.extra1		= SYSCTL_ZERO,
2939 	},
2940 	{
2941 		.procname	= "watermark_boost_factor",
2942 		.data		= &watermark_boost_factor,
2943 		.maxlen		= sizeof(watermark_boost_factor),
2944 		.mode		= 0644,
2945 		.proc_handler	= proc_dointvec_minmax,
2946 		.extra1		= SYSCTL_ZERO,
2947 	},
2948 	{
2949 		.procname	= "watermark_scale_factor",
2950 		.data		= &watermark_scale_factor,
2951 		.maxlen		= sizeof(watermark_scale_factor),
2952 		.mode		= 0644,
2953 		.proc_handler	= watermark_scale_factor_sysctl_handler,
2954 		.extra1		= SYSCTL_ONE,
2955 		.extra2		= &one_thousand,
2956 	},
2957 	{
2958 		.procname	= "extra_free_kbytes",
2959 		.data		= &extra_free_kbytes,
2960 		.maxlen		= sizeof(extra_free_kbytes),
2961 		.mode		= 0644,
2962 		.proc_handler	= min_free_kbytes_sysctl_handler,
2963 		.extra1		= SYSCTL_ZERO,
2964 	},
2965 	{
2966 		.procname	= "percpu_pagelist_fraction",
2967 		.data		= &percpu_pagelist_fraction,
2968 		.maxlen		= sizeof(percpu_pagelist_fraction),
2969 		.mode		= 0644,
2970 		.proc_handler	= percpu_pagelist_fraction_sysctl_handler,
2971 		.extra1		= SYSCTL_ZERO,
2972 	},
2973 	{
2974 		.procname	= "page_lock_unfairness",
2975 		.data		= &sysctl_page_lock_unfairness,
2976 		.maxlen		= sizeof(sysctl_page_lock_unfairness),
2977 		.mode		= 0644,
2978 		.proc_handler	= proc_dointvec_minmax,
2979 		.extra1		= SYSCTL_ZERO,
2980 	},
2981 #ifdef CONFIG_MMU
2982 	{
2983 		.procname	= "max_map_count",
2984 		.data		= &sysctl_max_map_count,
2985 		.maxlen		= sizeof(sysctl_max_map_count),
2986 		.mode		= 0644,
2987 		.proc_handler	= proc_dointvec_minmax,
2988 		.extra1		= SYSCTL_ZERO,
2989 	},
2990 #else
2991 	{
2992 		.procname	= "nr_trim_pages",
2993 		.data		= &sysctl_nr_trim_pages,
2994 		.maxlen		= sizeof(sysctl_nr_trim_pages),
2995 		.mode		= 0644,
2996 		.proc_handler	= proc_dointvec_minmax,
2997 		.extra1		= SYSCTL_ZERO,
2998 	},
2999 #endif
3000 	{
3001 		.procname	= "laptop_mode",
3002 		.data		= &laptop_mode,
3003 		.maxlen		= sizeof(laptop_mode),
3004 		.mode		= 0644,
3005 		.proc_handler	= proc_dointvec_jiffies,
3006 	},
3007 	{
3008 		.procname	= "block_dump",
3009 		.data		= &block_dump,
3010 		.maxlen		= sizeof(block_dump),
3011 		.mode		= 0644,
3012 		.proc_handler	= proc_dointvec_minmax,
3013 		.extra1		= SYSCTL_ZERO,
3014 	},
3015 	{
3016 		.procname	= "vfs_cache_pressure",
3017 		.data		= &sysctl_vfs_cache_pressure,
3018 		.maxlen		= sizeof(sysctl_vfs_cache_pressure),
3019 		.mode		= 0644,
3020 		.proc_handler	= proc_dointvec_minmax,
3021 		.extra1		= SYSCTL_ZERO,
3022 	},
3023 #if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
3024     defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
3025 	{
3026 		.procname	= "legacy_va_layout",
3027 		.data		= &sysctl_legacy_va_layout,
3028 		.maxlen		= sizeof(sysctl_legacy_va_layout),
3029 		.mode		= 0644,
3030 		.proc_handler	= proc_dointvec_minmax,
3031 		.extra1		= SYSCTL_ZERO,
3032 	},
3033 #endif
3034 #ifdef CONFIG_NUMA
3035 	{
3036 		.procname	= "zone_reclaim_mode",
3037 		.data		= &node_reclaim_mode,
3038 		.maxlen		= sizeof(node_reclaim_mode),
3039 		.mode		= 0644,
3040 		.proc_handler	= proc_dointvec_minmax,
3041 		.extra1		= SYSCTL_ZERO,
3042 	},
3043 	{
3044 		.procname	= "min_unmapped_ratio",
3045 		.data		= &sysctl_min_unmapped_ratio,
3046 		.maxlen		= sizeof(sysctl_min_unmapped_ratio),
3047 		.mode		= 0644,
3048 		.proc_handler	= sysctl_min_unmapped_ratio_sysctl_handler,
3049 		.extra1		= SYSCTL_ZERO,
3050 		.extra2		= &one_hundred,
3051 	},
3052 	{
3053 		.procname	= "min_slab_ratio",
3054 		.data		= &sysctl_min_slab_ratio,
3055 		.maxlen		= sizeof(sysctl_min_slab_ratio),
3056 		.mode		= 0644,
3057 		.proc_handler	= sysctl_min_slab_ratio_sysctl_handler,
3058 		.extra1		= SYSCTL_ZERO,
3059 		.extra2		= &one_hundred,
3060 	},
3061 #endif
3062 #ifdef CONFIG_SMP
3063 	{
3064 		.procname	= "stat_interval",
3065 		.data		= &sysctl_stat_interval,
3066 		.maxlen		= sizeof(sysctl_stat_interval),
3067 		.mode		= 0644,
3068 		.proc_handler	= proc_dointvec_jiffies,
3069 	},
3070 	{
3071 		.procname	= "stat_refresh",
3072 		.data		= NULL,
3073 		.maxlen		= 0,
3074 		.mode		= 0600,
3075 		.proc_handler	= vmstat_refresh,
3076 	},
3077 #endif
3078 #ifdef CONFIG_MMU
3079 	{
3080 		.procname	= "mmap_min_addr",
3081 		.data		= &dac_mmap_min_addr,
3082 		.maxlen		= sizeof(unsigned long),
3083 		.mode		= 0644,
3084 		.proc_handler	= mmap_min_addr_handler,
3085 	},
3086 #endif
3087 #ifdef CONFIG_NUMA
3088 	{
3089 		.procname	= "numa_zonelist_order",
3090 		.data		= &numa_zonelist_order,
3091 		.maxlen		= NUMA_ZONELIST_ORDER_LEN,
3092 		.mode		= 0644,
3093 		.proc_handler	= numa_zonelist_order_handler,
3094 	},
3095 #endif
3096 #if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
3097    (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
3098 	{
3099 		.procname	= "vdso_enabled",
3100 #ifdef CONFIG_X86_32
3101 		.data		= &vdso32_enabled,
3102 		.maxlen		= sizeof(vdso32_enabled),
3103 #else
3104 		.data		= &vdso_enabled,
3105 		.maxlen		= sizeof(vdso_enabled),
3106 #endif
3107 		.mode		= 0644,
3108 		.proc_handler	= proc_dointvec,
3109 		.extra1		= SYSCTL_ZERO,
3110 	},
3111 #endif
3112 #ifdef CONFIG_HIGHMEM
3113 	{
3114 		.procname	= "highmem_is_dirtyable",
3115 		.data		= &vm_highmem_is_dirtyable,
3116 		.maxlen		= sizeof(vm_highmem_is_dirtyable),
3117 		.mode		= 0644,
3118 		.proc_handler	= proc_dointvec_minmax,
3119 		.extra1		= SYSCTL_ZERO,
3120 		.extra2		= SYSCTL_ONE,
3121 	},
3122 #endif
3123 #ifdef CONFIG_MEMORY_FAILURE
3124 	{
3125 		.procname	= "memory_failure_early_kill",
3126 		.data		= &sysctl_memory_failure_early_kill,
3127 		.maxlen		= sizeof(sysctl_memory_failure_early_kill),
3128 		.mode		= 0644,
3129 		.proc_handler	= proc_dointvec_minmax,
3130 		.extra1		= SYSCTL_ZERO,
3131 		.extra2		= SYSCTL_ONE,
3132 	},
3133 	{
3134 		.procname	= "memory_failure_recovery",
3135 		.data		= &sysctl_memory_failure_recovery,
3136 		.maxlen		= sizeof(sysctl_memory_failure_recovery),
3137 		.mode		= 0644,
3138 		.proc_handler	= proc_dointvec_minmax,
3139 		.extra1		= SYSCTL_ZERO,
3140 		.extra2		= SYSCTL_ONE,
3141 	},
3142 #endif
3143 	{
3144 		.procname	= "user_reserve_kbytes",
3145 		.data		= &sysctl_user_reserve_kbytes,
3146 		.maxlen		= sizeof(sysctl_user_reserve_kbytes),
3147 		.mode		= 0644,
3148 		.proc_handler	= proc_doulongvec_minmax,
3149 	},
3150 	{
3151 		.procname	= "admin_reserve_kbytes",
3152 		.data		= &sysctl_admin_reserve_kbytes,
3153 		.maxlen		= sizeof(sysctl_admin_reserve_kbytes),
3154 		.mode		= 0644,
3155 		.proc_handler	= proc_doulongvec_minmax,
3156 	},
3157 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_BITS
3158 	{
3159 		.procname	= "mmap_rnd_bits",
3160 		.data		= &mmap_rnd_bits,
3161 		.maxlen		= sizeof(mmap_rnd_bits),
3162 		.mode		= 0600,
3163 		.proc_handler	= proc_dointvec_minmax,
3164 		.extra1		= (void *)&mmap_rnd_bits_min,
3165 		.extra2		= (void *)&mmap_rnd_bits_max,
3166 	},
3167 #endif
3168 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_COMPAT_BITS
3169 	{
3170 		.procname	= "mmap_rnd_compat_bits",
3171 		.data		= &mmap_rnd_compat_bits,
3172 		.maxlen		= sizeof(mmap_rnd_compat_bits),
3173 		.mode		= 0600,
3174 		.proc_handler	= proc_dointvec_minmax,
3175 		.extra1		= (void *)&mmap_rnd_compat_bits_min,
3176 		.extra2		= (void *)&mmap_rnd_compat_bits_max,
3177 	},
3178 #endif
3179 #ifdef CONFIG_USERFAULTFD
3180 	{
3181 		.procname	= "unprivileged_userfaultfd",
3182 		.data		= &sysctl_unprivileged_userfaultfd,
3183 		.maxlen		= sizeof(sysctl_unprivileged_userfaultfd),
3184 		.mode		= 0644,
3185 		.proc_handler	= proc_dointvec_minmax,
3186 		.extra1		= SYSCTL_ZERO,
3187 		.extra2		= SYSCTL_ONE,
3188 	},
3189 #endif
3190 	{ }
3191 };
3192 
3193 static struct ctl_table fs_table[] = {
3194 	{
3195 		.procname	= "inode-nr",
3196 		.data		= &inodes_stat,
3197 		.maxlen		= 2*sizeof(long),
3198 		.mode		= 0444,
3199 		.proc_handler	= proc_nr_inodes,
3200 	},
3201 	{
3202 		.procname	= "inode-state",
3203 		.data		= &inodes_stat,
3204 		.maxlen		= 7*sizeof(long),
3205 		.mode		= 0444,
3206 		.proc_handler	= proc_nr_inodes,
3207 	},
3208 	{
3209 		.procname	= "file-nr",
3210 		.data		= &files_stat,
3211 		.maxlen		= sizeof(files_stat),
3212 		.mode		= 0444,
3213 		.proc_handler	= proc_nr_files,
3214 	},
3215 	{
3216 		.procname	= "file-max",
3217 		.data		= &files_stat.max_files,
3218 		.maxlen		= sizeof(files_stat.max_files),
3219 		.mode		= 0644,
3220 		.proc_handler	= proc_doulongvec_minmax,
3221 		.extra1		= &zero_ul,
3222 		.extra2		= &long_max,
3223 	},
3224 	{
3225 		.procname	= "nr_open",
3226 		.data		= &sysctl_nr_open,
3227 		.maxlen		= sizeof(unsigned int),
3228 		.mode		= 0644,
3229 		.proc_handler	= proc_dointvec_minmax,
3230 		.extra1		= &sysctl_nr_open_min,
3231 		.extra2		= &sysctl_nr_open_max,
3232 	},
3233 	{
3234 		.procname	= "dentry-state",
3235 		.data		= &dentry_stat,
3236 		.maxlen		= 6*sizeof(long),
3237 		.mode		= 0444,
3238 		.proc_handler	= proc_nr_dentry,
3239 	},
3240 	{
3241 		.procname	= "overflowuid",
3242 		.data		= &fs_overflowuid,
3243 		.maxlen		= sizeof(int),
3244 		.mode		= 0644,
3245 		.proc_handler	= proc_dointvec_minmax,
3246 		.extra1		= &minolduid,
3247 		.extra2		= &maxolduid,
3248 	},
3249 	{
3250 		.procname	= "overflowgid",
3251 		.data		= &fs_overflowgid,
3252 		.maxlen		= sizeof(int),
3253 		.mode		= 0644,
3254 		.proc_handler	= proc_dointvec_minmax,
3255 		.extra1		= &minolduid,
3256 		.extra2		= &maxolduid,
3257 	},
3258 #ifdef CONFIG_FILE_LOCKING
3259 	{
3260 		.procname	= "leases-enable",
3261 		.data		= &leases_enable,
3262 		.maxlen		= sizeof(int),
3263 		.mode		= 0644,
3264 		.proc_handler	= proc_dointvec,
3265 	},
3266 #endif
3267 #ifdef CONFIG_DNOTIFY
3268 	{
3269 		.procname	= "dir-notify-enable",
3270 		.data		= &dir_notify_enable,
3271 		.maxlen		= sizeof(int),
3272 		.mode		= 0644,
3273 		.proc_handler	= proc_dointvec,
3274 	},
3275 #endif
3276 #ifdef CONFIG_MMU
3277 #ifdef CONFIG_FILE_LOCKING
3278 	{
3279 		.procname	= "lease-break-time",
3280 		.data		= &lease_break_time,
3281 		.maxlen		= sizeof(int),
3282 		.mode		= 0644,
3283 		.proc_handler	= proc_dointvec,
3284 	},
3285 #endif
3286 #ifdef CONFIG_AIO
3287 	{
3288 		.procname	= "aio-nr",
3289 		.data		= &aio_nr,
3290 		.maxlen		= sizeof(aio_nr),
3291 		.mode		= 0444,
3292 		.proc_handler	= proc_doulongvec_minmax,
3293 	},
3294 	{
3295 		.procname	= "aio-max-nr",
3296 		.data		= &aio_max_nr,
3297 		.maxlen		= sizeof(aio_max_nr),
3298 		.mode		= 0644,
3299 		.proc_handler	= proc_doulongvec_minmax,
3300 	},
3301 #endif /* CONFIG_AIO */
3302 #ifdef CONFIG_INOTIFY_USER
3303 	{
3304 		.procname	= "inotify",
3305 		.mode		= 0555,
3306 		.child		= inotify_table,
3307 	},
3308 #endif
3309 #ifdef CONFIG_EPOLL
3310 	{
3311 		.procname	= "epoll",
3312 		.mode		= 0555,
3313 		.child		= epoll_table,
3314 	},
3315 #endif
3316 #endif
3317 	{
3318 		.procname	= "protected_symlinks",
3319 		.data		= &sysctl_protected_symlinks,
3320 		.maxlen		= sizeof(int),
3321 		.mode		= 0600,
3322 		.proc_handler	= proc_dointvec_minmax,
3323 		.extra1		= SYSCTL_ZERO,
3324 		.extra2		= SYSCTL_ONE,
3325 	},
3326 	{
3327 		.procname	= "protected_hardlinks",
3328 		.data		= &sysctl_protected_hardlinks,
3329 		.maxlen		= sizeof(int),
3330 		.mode		= 0600,
3331 		.proc_handler	= proc_dointvec_minmax,
3332 		.extra1		= SYSCTL_ZERO,
3333 		.extra2		= SYSCTL_ONE,
3334 	},
3335 	{
3336 		.procname	= "protected_fifos",
3337 		.data		= &sysctl_protected_fifos,
3338 		.maxlen		= sizeof(int),
3339 		.mode		= 0600,
3340 		.proc_handler	= proc_dointvec_minmax,
3341 		.extra1		= SYSCTL_ZERO,
3342 		.extra2		= &two,
3343 	},
3344 	{
3345 		.procname	= "protected_regular",
3346 		.data		= &sysctl_protected_regular,
3347 		.maxlen		= sizeof(int),
3348 		.mode		= 0600,
3349 		.proc_handler	= proc_dointvec_minmax,
3350 		.extra1		= SYSCTL_ZERO,
3351 		.extra2		= &two,
3352 	},
3353 	{
3354 		.procname	= "suid_dumpable",
3355 		.data		= &suid_dumpable,
3356 		.maxlen		= sizeof(int),
3357 		.mode		= 0644,
3358 		.proc_handler	= proc_dointvec_minmax_coredump,
3359 		.extra1		= SYSCTL_ZERO,
3360 		.extra2		= &two,
3361 	},
3362 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
3363 	{
3364 		.procname	= "binfmt_misc",
3365 		.mode		= 0555,
3366 		.child		= sysctl_mount_point,
3367 	},
3368 #endif
3369 	{
3370 		.procname	= "pipe-max-size",
3371 		.data		= &pipe_max_size,
3372 		.maxlen		= sizeof(pipe_max_size),
3373 		.mode		= 0644,
3374 		.proc_handler	= proc_dopipe_max_size,
3375 	},
3376 	{
3377 		.procname	= "pipe-user-pages-hard",
3378 		.data		= &pipe_user_pages_hard,
3379 		.maxlen		= sizeof(pipe_user_pages_hard),
3380 		.mode		= 0644,
3381 		.proc_handler	= proc_doulongvec_minmax,
3382 	},
3383 	{
3384 		.procname	= "pipe-user-pages-soft",
3385 		.data		= &pipe_user_pages_soft,
3386 		.maxlen		= sizeof(pipe_user_pages_soft),
3387 		.mode		= 0644,
3388 		.proc_handler	= proc_doulongvec_minmax,
3389 	},
3390 	{
3391 		.procname	= "mount-max",
3392 		.data		= &sysctl_mount_max,
3393 		.maxlen		= sizeof(unsigned int),
3394 		.mode		= 0644,
3395 		.proc_handler	= proc_dointvec_minmax,
3396 		.extra1		= SYSCTL_ONE,
3397 	},
3398 	{ }
3399 };
3400 
3401 static struct ctl_table debug_table[] = {
3402 #ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
3403 	{
3404 		.procname	= "exception-trace",
3405 		.data		= &show_unhandled_signals,
3406 		.maxlen		= sizeof(int),
3407 		.mode		= 0644,
3408 		.proc_handler	= proc_dointvec
3409 	},
3410 #endif
3411 #if defined(CONFIG_OPTPROBES)
3412 	{
3413 		.procname	= "kprobes-optimization",
3414 		.data		= &sysctl_kprobes_optimization,
3415 		.maxlen		= sizeof(int),
3416 		.mode		= 0644,
3417 		.proc_handler	= proc_kprobes_optimization_handler,
3418 		.extra1		= SYSCTL_ZERO,
3419 		.extra2		= SYSCTL_ONE,
3420 	},
3421 #endif
3422 	{ }
3423 };
3424 
3425 static struct ctl_table dev_table[] = {
3426 	{ }
3427 };
3428 
3429 static struct ctl_table sysctl_base_table[] = {
3430 	{
3431 		.procname	= "kernel",
3432 		.mode		= 0555,
3433 		.child		= kern_table,
3434 	},
3435 	{
3436 		.procname	= "vm",
3437 		.mode		= 0555,
3438 		.child		= vm_table,
3439 	},
3440 	{
3441 		.procname	= "fs",
3442 		.mode		= 0555,
3443 		.child		= fs_table,
3444 	},
3445 	{
3446 		.procname	= "debug",
3447 		.mode		= 0555,
3448 		.child		= debug_table,
3449 	},
3450 	{
3451 		.procname	= "dev",
3452 		.mode		= 0555,
3453 		.child		= dev_table,
3454 	},
3455 	{ }
3456 };
3457 
sysctl_init(void)3458 int __init sysctl_init(void)
3459 {
3460 	struct ctl_table_header *hdr;
3461 
3462 	hdr = register_sysctl_table(sysctl_base_table);
3463 	kmemleak_not_leak(hdr);
3464 	return 0;
3465 }
3466 #endif /* CONFIG_SYSCTL */
3467 /*
3468  * No sense putting this after each symbol definition, twice,
3469  * exception granted :-)
3470  */
3471 EXPORT_SYMBOL(proc_dointvec);
3472 EXPORT_SYMBOL(proc_douintvec);
3473 EXPORT_SYMBOL(proc_dointvec_jiffies);
3474 EXPORT_SYMBOL(proc_dointvec_minmax);
3475 EXPORT_SYMBOL_GPL(proc_douintvec_minmax);
3476 EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
3477 EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
3478 EXPORT_SYMBOL(proc_dostring);
3479 EXPORT_SYMBOL(proc_doulongvec_minmax);
3480 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
3481 EXPORT_SYMBOL(proc_do_large_bitmap);
3482