Linux Audio

Check our new training course

Loading...
v5.4
   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
  72#include "../lib/kstrtox.h"
  73
  74#include <linux/uaccess.h>
  75#include <asm/processor.h>
  76
  77#ifdef CONFIG_X86
  78#include <asm/nmi.h>
  79#include <asm/stacktrace.h>
  80#include <asm/io.h>
  81#endif
  82#ifdef CONFIG_SPARC
  83#include <asm/setup.h>
  84#endif
  85#ifdef CONFIG_BSD_PROCESS_ACCT
  86#include <linux/acct.h>
  87#endif
  88#ifdef CONFIG_RT_MUTEXES
  89#include <linux/rtmutex.h>
  90#endif
  91#if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
  92#include <linux/lockdep.h>
  93#endif
  94#ifdef CONFIG_CHR_DEV_SG
  95#include <scsi/sg.h>
  96#endif
  97#ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
  98#include <linux/stackleak.h>
  99#endif
 100#ifdef CONFIG_LOCKUP_DETECTOR
 101#include <linux/nmi.h>
 102#endif
 103
 104#if defined(CONFIG_SYSCTL)
 105
 106/* External variables not in a header file. */
 107extern int suid_dumpable;
 108#ifdef CONFIG_COREDUMP
 109extern int core_uses_pid;
 110extern char core_pattern[];
 111extern unsigned int core_pipe_limit;
 112#endif
 113extern int pid_max;
 114extern int pid_max_min, pid_max_max;
 115extern int percpu_pagelist_fraction;
 116extern int latencytop_enabled;
 117extern unsigned int sysctl_nr_open_min, sysctl_nr_open_max;
 118#ifndef CONFIG_MMU
 119extern int sysctl_nr_trim_pages;
 120#endif
 121
 122/* Constants used for minimum and  maximum */
 123#ifdef CONFIG_LOCKUP_DETECTOR
 124static int sixty = 60;
 125#endif
 126
 127static int __maybe_unused neg_one = -1;
 128static int __maybe_unused two = 2;
 129static int __maybe_unused four = 4;
 130static unsigned long zero_ul;
 131static unsigned long one_ul = 1;
 132static unsigned long long_max = LONG_MAX;
 133static int one_hundred = 100;
 
 134static int one_thousand = 1000;
 135#ifdef CONFIG_PRINTK
 136static int ten_thousand = 10000;
 137#endif
 138#ifdef CONFIG_PERF_EVENTS
 139static int six_hundred_forty_kb = 640 * 1024;
 140#endif
 141
 142/* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
 143static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
 144
 145/* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
 146static int maxolduid = 65535;
 147static int minolduid;
 148
 149static int ngroups_max = NGROUPS_MAX;
 150static const int cap_last_cap = CAP_LAST_CAP;
 151
 152/*
 153 * This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
 154 * and hung_task_check_interval_secs
 155 */
 156#ifdef CONFIG_DETECT_HUNG_TASK
 157static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
 158#endif
 159
 160#ifdef CONFIG_INOTIFY_USER
 161#include <linux/inotify.h>
 162#endif
 163#ifdef CONFIG_SPARC
 164#endif
 165
 166#ifdef CONFIG_PARISC
 167extern int pwrsw_enabled;
 168#endif
 169
 170#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
 171extern int unaligned_enabled;
 172#endif
 173
 174#ifdef CONFIG_IA64
 175extern int unaligned_dump_stack;
 176#endif
 177
 178#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
 179extern int no_unaligned_warning;
 180#endif
 181
 182#ifdef CONFIG_PROC_SYSCTL
 183
 184/**
 185 * enum sysctl_writes_mode - supported sysctl write modes
 186 *
 187 * @SYSCTL_WRITES_LEGACY: each write syscall must fully contain the sysctl value
 188 *	to be written, and multiple writes on the same sysctl file descriptor
 189 *	will rewrite the sysctl value, regardless of file position. No warning
 190 *	is issued when the initial position is not 0.
 191 * @SYSCTL_WRITES_WARN: same as above but warn when the initial file position is
 192 *	not 0.
 193 * @SYSCTL_WRITES_STRICT: writes to numeric sysctl entries must always be at
 194 *	file position 0 and the value must be fully contained in the buffer
 195 *	sent to the write syscall. If dealing with strings respect the file
 196 *	position, but restrict this to the max length of the buffer, anything
 197 *	passed the max length will be ignored. Multiple writes will append
 198 *	to the buffer.
 199 *
 200 * These write modes control how current file position affects the behavior of
 201 * updating sysctl values through the proc interface on each write.
 202 */
 203enum sysctl_writes_mode {
 204	SYSCTL_WRITES_LEGACY		= -1,
 205	SYSCTL_WRITES_WARN		= 0,
 206	SYSCTL_WRITES_STRICT		= 1,
 207};
 208
 209static enum sysctl_writes_mode sysctl_writes_strict = SYSCTL_WRITES_STRICT;
 
 210
 211static int proc_do_cad_pid(struct ctl_table *table, int write,
 212		  void __user *buffer, size_t *lenp, loff_t *ppos);
 213static int proc_taint(struct ctl_table *table, int write,
 214			       void __user *buffer, size_t *lenp, loff_t *ppos);
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 215#endif
 216
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 217#ifdef CONFIG_PRINTK
 218static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
 219				void __user *buffer, size_t *lenp, loff_t *ppos);
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 220#endif
 
 221
 222static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
 223		void __user *buffer, size_t *lenp, loff_t *ppos);
 
 
 
 
 
 
 
 224#ifdef CONFIG_COREDUMP
 225static int proc_dostring_coredump(struct ctl_table *table, int write,
 226		void __user *buffer, size_t *lenp, loff_t *ppos);
 
 
 
 
 
 
 227#endif
 228static int proc_dopipe_max_size(struct ctl_table *table, int write,
 229		void __user *buffer, size_t *lenp, loff_t *ppos);
 230
 231#ifdef CONFIG_MAGIC_SYSRQ
 232/* Note: sysrq code uses its own private copy */
 233static int __sysrq_enabled = CONFIG_MAGIC_SYSRQ_DEFAULT_ENABLE;
 234
 235static int sysrq_sysctl_handler(struct ctl_table *table, int write,
 236				void __user *buffer, size_t *lenp,
 237				loff_t *ppos)
 238{
 239	int error;
 
 
 240
 241	error = proc_dointvec(table, write, buffer, lenp, ppos);
 242	if (error)
 243		return error;
 
 244
 245	if (write)
 246		sysrq_toggle_support(__sysrq_enabled);
 247
 248	return 0;
 249}
 250
 251#endif
 252
 253static struct ctl_table kern_table[];
 254static struct ctl_table vm_table[];
 255static struct ctl_table fs_table[];
 256static struct ctl_table debug_table[];
 257static struct ctl_table dev_table[];
 258extern struct ctl_table random_table[];
 259#ifdef CONFIG_EPOLL
 260extern struct ctl_table epoll_table[];
 261#endif
 262
 263#ifdef CONFIG_FW_LOADER_USER_HELPER
 264extern struct ctl_table firmware_config_table[];
 265#endif
 
 266
 267#if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
 268    defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
 269int sysctl_legacy_va_layout;
 270#endif
 
 271
 272/* The default sysctl tables: */
 
 
 273
 274static struct ctl_table sysctl_base_table[] = {
 275	{
 276		.procname	= "kernel",
 277		.mode		= 0555,
 278		.child		= kern_table,
 279	},
 280	{
 281		.procname	= "vm",
 282		.mode		= 0555,
 283		.child		= vm_table,
 284	},
 285	{
 286		.procname	= "fs",
 287		.mode		= 0555,
 288		.child		= fs_table,
 289	},
 290	{
 291		.procname	= "debug",
 292		.mode		= 0555,
 293		.child		= debug_table,
 294	},
 295	{
 296		.procname	= "dev",
 297		.mode		= 0555,
 298		.child		= dev_table,
 299	},
 300	{ }
 301};
 302
 303#ifdef CONFIG_SCHED_DEBUG
 304static int min_sched_granularity_ns = 100000;		/* 100 usecs */
 305static int max_sched_granularity_ns = NSEC_PER_SEC;	/* 1 second */
 306static int min_wakeup_granularity_ns;			/* 0 usecs */
 307static int max_wakeup_granularity_ns = NSEC_PER_SEC;	/* 1 second */
 308#ifdef CONFIG_SMP
 309static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
 310static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
 311#endif /* CONFIG_SMP */
 312#endif /* CONFIG_SCHED_DEBUG */
 313
 314#ifdef CONFIG_COMPACTION
 315static int min_extfrag_threshold;
 316static int max_extfrag_threshold = 1000;
 317#endif
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 318
 319static struct ctl_table kern_table[] = {
 320	{
 321		.procname	= "sched_child_runs_first",
 322		.data		= &sysctl_sched_child_runs_first,
 323		.maxlen		= sizeof(unsigned int),
 324		.mode		= 0644,
 325		.proc_handler	= proc_dointvec,
 326	},
 327#ifdef CONFIG_SCHED_DEBUG
 328	{
 329		.procname	= "sched_min_granularity_ns",
 330		.data		= &sysctl_sched_min_granularity,
 331		.maxlen		= sizeof(unsigned int),
 332		.mode		= 0644,
 333		.proc_handler	= sched_proc_update_handler,
 334		.extra1		= &min_sched_granularity_ns,
 335		.extra2		= &max_sched_granularity_ns,
 336	},
 337	{
 338		.procname	= "sched_latency_ns",
 339		.data		= &sysctl_sched_latency,
 340		.maxlen		= sizeof(unsigned int),
 341		.mode		= 0644,
 342		.proc_handler	= sched_proc_update_handler,
 343		.extra1		= &min_sched_granularity_ns,
 344		.extra2		= &max_sched_granularity_ns,
 345	},
 346	{
 347		.procname	= "sched_wakeup_granularity_ns",
 348		.data		= &sysctl_sched_wakeup_granularity,
 349		.maxlen		= sizeof(unsigned int),
 350		.mode		= 0644,
 351		.proc_handler	= sched_proc_update_handler,
 352		.extra1		= &min_wakeup_granularity_ns,
 353		.extra2		= &max_wakeup_granularity_ns,
 354	},
 355#ifdef CONFIG_SMP
 356	{
 357		.procname	= "sched_tunable_scaling",
 358		.data		= &sysctl_sched_tunable_scaling,
 359		.maxlen		= sizeof(enum sched_tunable_scaling),
 360		.mode		= 0644,
 361		.proc_handler	= sched_proc_update_handler,
 362		.extra1		= &min_sched_tunable_scaling,
 363		.extra2		= &max_sched_tunable_scaling,
 364	},
 365	{
 366		.procname	= "sched_migration_cost_ns",
 367		.data		= &sysctl_sched_migration_cost,
 368		.maxlen		= sizeof(unsigned int),
 369		.mode		= 0644,
 370		.proc_handler	= proc_dointvec,
 371	},
 372	{
 373		.procname	= "sched_nr_migrate",
 374		.data		= &sysctl_sched_nr_migrate,
 375		.maxlen		= sizeof(unsigned int),
 376		.mode		= 0644,
 377		.proc_handler	= proc_dointvec,
 378	},
 379#ifdef CONFIG_SCHEDSTATS
 380	{
 381		.procname	= "sched_schedstats",
 382		.data		= NULL,
 383		.maxlen		= sizeof(unsigned int),
 384		.mode		= 0644,
 385		.proc_handler	= sysctl_schedstats,
 386		.extra1		= SYSCTL_ZERO,
 387		.extra2		= SYSCTL_ONE,
 388	},
 389#endif /* CONFIG_SCHEDSTATS */
 390#endif /* CONFIG_SMP */
 391#ifdef CONFIG_NUMA_BALANCING
 392	{
 393		.procname	= "numa_balancing_scan_delay_ms",
 394		.data		= &sysctl_numa_balancing_scan_delay,
 395		.maxlen		= sizeof(unsigned int),
 396		.mode		= 0644,
 397		.proc_handler	= proc_dointvec,
 398	},
 399	{
 400		.procname	= "numa_balancing_scan_period_min_ms",
 401		.data		= &sysctl_numa_balancing_scan_period_min,
 402		.maxlen		= sizeof(unsigned int),
 403		.mode		= 0644,
 404		.proc_handler	= proc_dointvec,
 405	},
 406	{
 407		.procname	= "numa_balancing_scan_period_max_ms",
 408		.data		= &sysctl_numa_balancing_scan_period_max,
 409		.maxlen		= sizeof(unsigned int),
 410		.mode		= 0644,
 411		.proc_handler	= proc_dointvec,
 412	},
 413	{
 414		.procname	= "numa_balancing_scan_size_mb",
 415		.data		= &sysctl_numa_balancing_scan_size,
 416		.maxlen		= sizeof(unsigned int),
 417		.mode		= 0644,
 418		.proc_handler	= proc_dointvec_minmax,
 419		.extra1		= SYSCTL_ONE,
 420	},
 421	{
 422		.procname	= "numa_balancing",
 423		.data		= NULL, /* filled in by handler */
 424		.maxlen		= sizeof(unsigned int),
 425		.mode		= 0644,
 426		.proc_handler	= sysctl_numa_balancing,
 427		.extra1		= SYSCTL_ZERO,
 428		.extra2		= SYSCTL_ONE,
 429	},
 430#endif /* CONFIG_NUMA_BALANCING */
 431#endif /* CONFIG_SCHED_DEBUG */
 432	{
 433		.procname	= "sched_rt_period_us",
 434		.data		= &sysctl_sched_rt_period,
 435		.maxlen		= sizeof(unsigned int),
 436		.mode		= 0644,
 437		.proc_handler	= sched_rt_handler,
 438	},
 439	{
 440		.procname	= "sched_rt_runtime_us",
 441		.data		= &sysctl_sched_rt_runtime,
 442		.maxlen		= sizeof(int),
 443		.mode		= 0644,
 444		.proc_handler	= sched_rt_handler,
 445	},
 446	{
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 447		.procname	= "sched_rr_timeslice_ms",
 448		.data		= &sysctl_sched_rr_timeslice,
 449		.maxlen		= sizeof(int),
 450		.mode		= 0644,
 451		.proc_handler	= sched_rr_handler,
 452	},
 453#ifdef CONFIG_UCLAMP_TASK
 454	{
 455		.procname	= "sched_util_clamp_min",
 456		.data		= &sysctl_sched_uclamp_util_min,
 457		.maxlen		= sizeof(unsigned int),
 458		.mode		= 0644,
 459		.proc_handler	= sysctl_sched_uclamp_handler,
 460	},
 461	{
 462		.procname	= "sched_util_clamp_max",
 463		.data		= &sysctl_sched_uclamp_util_max,
 464		.maxlen		= sizeof(unsigned int),
 465		.mode		= 0644,
 466		.proc_handler	= sysctl_sched_uclamp_handler,
 467	},
 
 
 
 
 
 
 
 468#endif
 469#ifdef CONFIG_SCHED_AUTOGROUP
 470	{
 471		.procname	= "sched_autogroup_enabled",
 472		.data		= &sysctl_sched_autogroup_enabled,
 473		.maxlen		= sizeof(unsigned int),
 474		.mode		= 0644,
 475		.proc_handler	= proc_dointvec_minmax,
 476		.extra1		= SYSCTL_ZERO,
 477		.extra2		= SYSCTL_ONE,
 478	},
 479#endif
 480#ifdef CONFIG_CFS_BANDWIDTH
 481	{
 482		.procname	= "sched_cfs_bandwidth_slice_us",
 483		.data		= &sysctl_sched_cfs_bandwidth_slice,
 484		.maxlen		= sizeof(unsigned int),
 485		.mode		= 0644,
 486		.proc_handler	= proc_dointvec_minmax,
 487		.extra1		= SYSCTL_ONE,
 488	},
 489#endif
 490#if defined(CONFIG_ENERGY_MODEL) && defined(CONFIG_CPU_FREQ_GOV_SCHEDUTIL)
 491	{
 492		.procname	= "sched_energy_aware",
 493		.data		= &sysctl_sched_energy_aware,
 494		.maxlen		= sizeof(unsigned int),
 495		.mode		= 0644,
 496		.proc_handler	= sched_energy_aware_handler,
 497		.extra1		= SYSCTL_ZERO,
 498		.extra2		= SYSCTL_ONE,
 499	},
 500#endif
 501#ifdef CONFIG_PROVE_LOCKING
 502	{
 503		.procname	= "prove_locking",
 504		.data		= &prove_locking,
 505		.maxlen		= sizeof(int),
 506		.mode		= 0644,
 507		.proc_handler	= proc_dointvec,
 508	},
 509#endif
 510#ifdef CONFIG_LOCK_STAT
 511	{
 512		.procname	= "lock_stat",
 513		.data		= &lock_stat,
 514		.maxlen		= sizeof(int),
 515		.mode		= 0644,
 516		.proc_handler	= proc_dointvec,
 517	},
 518#endif
 519	{
 520		.procname	= "panic",
 521		.data		= &panic_timeout,
 522		.maxlen		= sizeof(int),
 523		.mode		= 0644,
 524		.proc_handler	= proc_dointvec,
 525	},
 526#ifdef CONFIG_COREDUMP
 527	{
 528		.procname	= "core_uses_pid",
 529		.data		= &core_uses_pid,
 530		.maxlen		= sizeof(int),
 531		.mode		= 0644,
 532		.proc_handler	= proc_dointvec,
 533	},
 534	{
 535		.procname	= "core_pattern",
 536		.data		= core_pattern,
 537		.maxlen		= CORENAME_MAX_SIZE,
 538		.mode		= 0644,
 539		.proc_handler	= proc_dostring_coredump,
 540	},
 541	{
 542		.procname	= "core_pipe_limit",
 543		.data		= &core_pipe_limit,
 544		.maxlen		= sizeof(unsigned int),
 545		.mode		= 0644,
 546		.proc_handler	= proc_dointvec,
 547	},
 548#endif
 549#ifdef CONFIG_PROC_SYSCTL
 550	{
 551		.procname	= "tainted",
 552		.maxlen 	= sizeof(long),
 553		.mode		= 0644,
 554		.proc_handler	= proc_taint,
 555	},
 556	{
 557		.procname	= "sysctl_writes_strict",
 558		.data		= &sysctl_writes_strict,
 559		.maxlen		= sizeof(int),
 560		.mode		= 0644,
 561		.proc_handler	= proc_dointvec_minmax,
 562		.extra1		= &neg_one,
 563		.extra2		= SYSCTL_ONE,
 564	},
 565#endif
 566#ifdef CONFIG_LATENCYTOP
 567	{
 568		.procname	= "latencytop",
 569		.data		= &latencytop_enabled,
 570		.maxlen		= sizeof(int),
 571		.mode		= 0644,
 572		.proc_handler	= sysctl_latencytop,
 573	},
 574#endif
 575#ifdef CONFIG_BLK_DEV_INITRD
 576	{
 577		.procname	= "real-root-dev",
 578		.data		= &real_root_dev,
 579		.maxlen		= sizeof(int),
 580		.mode		= 0644,
 581		.proc_handler	= proc_dointvec,
 582	},
 583#endif
 584	{
 585		.procname	= "print-fatal-signals",
 586		.data		= &print_fatal_signals,
 587		.maxlen		= sizeof(int),
 588		.mode		= 0644,
 589		.proc_handler	= proc_dointvec,
 590	},
 591#ifdef CONFIG_SPARC
 592	{
 593		.procname	= "reboot-cmd",
 594		.data		= reboot_command,
 595		.maxlen		= 256,
 596		.mode		= 0644,
 597		.proc_handler	= proc_dostring,
 598	},
 599	{
 600		.procname	= "stop-a",
 601		.data		= &stop_a_enabled,
 602		.maxlen		= sizeof (int),
 603		.mode		= 0644,
 604		.proc_handler	= proc_dointvec,
 605	},
 606	{
 607		.procname	= "scons-poweroff",
 608		.data		= &scons_pwroff,
 609		.maxlen		= sizeof (int),
 610		.mode		= 0644,
 611		.proc_handler	= proc_dointvec,
 612	},
 613#endif
 614#ifdef CONFIG_SPARC64
 615	{
 616		.procname	= "tsb-ratio",
 617		.data		= &sysctl_tsb_ratio,
 618		.maxlen		= sizeof (int),
 619		.mode		= 0644,
 620		.proc_handler	= proc_dointvec,
 621	},
 622#endif
 623#ifdef CONFIG_PARISC
 624	{
 625		.procname	= "soft-power",
 626		.data		= &pwrsw_enabled,
 627		.maxlen		= sizeof (int),
 628	 	.mode		= 0644,
 629		.proc_handler	= proc_dointvec,
 630	},
 631#endif
 632#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
 633	{
 634		.procname	= "unaligned-trap",
 635		.data		= &unaligned_enabled,
 636		.maxlen		= sizeof (int),
 637		.mode		= 0644,
 638		.proc_handler	= proc_dointvec,
 639	},
 640#endif
 641	{
 642		.procname	= "ctrl-alt-del",
 643		.data		= &C_A_D,
 644		.maxlen		= sizeof(int),
 645		.mode		= 0644,
 646		.proc_handler	= proc_dointvec,
 647	},
 648#ifdef CONFIG_FUNCTION_TRACER
 649	{
 650		.procname	= "ftrace_enabled",
 651		.data		= &ftrace_enabled,
 652		.maxlen		= sizeof(int),
 653		.mode		= 0644,
 654		.proc_handler	= ftrace_enable_sysctl,
 655	},
 656#endif
 657#ifdef CONFIG_STACK_TRACER
 658	{
 659		.procname	= "stack_tracer_enabled",
 660		.data		= &stack_tracer_enabled,
 661		.maxlen		= sizeof(int),
 662		.mode		= 0644,
 663		.proc_handler	= stack_trace_sysctl,
 664	},
 665#endif
 666#ifdef CONFIG_TRACING
 667	{
 668		.procname	= "ftrace_dump_on_oops",
 669		.data		= &ftrace_dump_on_oops,
 670		.maxlen		= sizeof(int),
 671		.mode		= 0644,
 672		.proc_handler	= proc_dointvec,
 673	},
 674	{
 675		.procname	= "traceoff_on_warning",
 676		.data		= &__disable_trace_on_warning,
 677		.maxlen		= sizeof(__disable_trace_on_warning),
 678		.mode		= 0644,
 679		.proc_handler	= proc_dointvec,
 680	},
 681	{
 682		.procname	= "tracepoint_printk",
 683		.data		= &tracepoint_printk,
 684		.maxlen		= sizeof(tracepoint_printk),
 685		.mode		= 0644,
 686		.proc_handler	= tracepoint_printk_sysctl,
 687	},
 688#endif
 689#ifdef CONFIG_KEXEC_CORE
 690	{
 691		.procname	= "kexec_load_disabled",
 692		.data		= &kexec_load_disabled,
 693		.maxlen		= sizeof(int),
 694		.mode		= 0644,
 695		/* only handle a transition from default "0" to "1" */
 696		.proc_handler	= proc_dointvec_minmax,
 697		.extra1		= SYSCTL_ONE,
 698		.extra2		= SYSCTL_ONE,
 699	},
 700#endif
 701#ifdef CONFIG_MODULES
 702	{
 703		.procname	= "modprobe",
 704		.data		= &modprobe_path,
 705		.maxlen		= KMOD_PATH_LEN,
 706		.mode		= 0644,
 707		.proc_handler	= proc_dostring,
 708	},
 709	{
 710		.procname	= "modules_disabled",
 711		.data		= &modules_disabled,
 712		.maxlen		= sizeof(int),
 713		.mode		= 0644,
 714		/* only handle a transition from default "0" to "1" */
 715		.proc_handler	= proc_dointvec_minmax,
 716		.extra1		= SYSCTL_ONE,
 717		.extra2		= SYSCTL_ONE,
 718	},
 719#endif
 720#ifdef CONFIG_UEVENT_HELPER
 721	{
 722		.procname	= "hotplug",
 723		.data		= &uevent_helper,
 724		.maxlen		= UEVENT_HELPER_PATH_LEN,
 725		.mode		= 0644,
 726		.proc_handler	= proc_dostring,
 727	},
 728#endif
 729#ifdef CONFIG_CHR_DEV_SG
 730	{
 731		.procname	= "sg-big-buff",
 732		.data		= &sg_big_buff,
 733		.maxlen		= sizeof (int),
 734		.mode		= 0444,
 735		.proc_handler	= proc_dointvec,
 736	},
 737#endif
 738#ifdef CONFIG_BSD_PROCESS_ACCT
 739	{
 740		.procname	= "acct",
 741		.data		= &acct_parm,
 742		.maxlen		= 3*sizeof(int),
 743		.mode		= 0644,
 744		.proc_handler	= proc_dointvec,
 745	},
 746#endif
 747#ifdef CONFIG_MAGIC_SYSRQ
 748	{
 749		.procname	= "sysrq",
 750		.data		= &__sysrq_enabled,
 751		.maxlen		= sizeof (int),
 752		.mode		= 0644,
 753		.proc_handler	= sysrq_sysctl_handler,
 754	},
 755#endif
 756#ifdef CONFIG_PROC_SYSCTL
 757	{
 758		.procname	= "cad_pid",
 759		.data		= NULL,
 760		.maxlen		= sizeof (int),
 761		.mode		= 0600,
 762		.proc_handler	= proc_do_cad_pid,
 763	},
 764#endif
 765	{
 766		.procname	= "threads-max",
 767		.data		= NULL,
 768		.maxlen		= sizeof(int),
 769		.mode		= 0644,
 770		.proc_handler	= sysctl_max_threads,
 771	},
 772	{
 773		.procname	= "random",
 774		.mode		= 0555,
 775		.child		= random_table,
 776	},
 777	{
 778		.procname	= "usermodehelper",
 779		.mode		= 0555,
 780		.child		= usermodehelper_table,
 781	},
 782#ifdef CONFIG_FW_LOADER_USER_HELPER
 783	{
 784		.procname	= "firmware_config",
 785		.mode		= 0555,
 786		.child		= firmware_config_table,
 787	},
 788#endif
 789	{
 790		.procname	= "overflowuid",
 791		.data		= &overflowuid,
 792		.maxlen		= sizeof(int),
 793		.mode		= 0644,
 794		.proc_handler	= proc_dointvec_minmax,
 795		.extra1		= &minolduid,
 796		.extra2		= &maxolduid,
 797	},
 798	{
 799		.procname	= "overflowgid",
 800		.data		= &overflowgid,
 801		.maxlen		= sizeof(int),
 802		.mode		= 0644,
 803		.proc_handler	= proc_dointvec_minmax,
 804		.extra1		= &minolduid,
 805		.extra2		= &maxolduid,
 806	},
 807#ifdef CONFIG_S390
 808#ifdef CONFIG_MATHEMU
 809	{
 810		.procname	= "ieee_emulation_warnings",
 811		.data		= &sysctl_ieee_emulation_warnings,
 812		.maxlen		= sizeof(int),
 813		.mode		= 0644,
 814		.proc_handler	= proc_dointvec,
 815	},
 816#endif
 
 817	{
 818		.procname	= "userprocess_debug",
 819		.data		= &show_unhandled_signals,
 820		.maxlen		= sizeof(int),
 821		.mode		= 0644,
 822		.proc_handler	= proc_dointvec,
 
 
 823	},
 824#endif
 825	{
 826		.procname	= "pid_max",
 827		.data		= &pid_max,
 828		.maxlen		= sizeof (int),
 829		.mode		= 0644,
 830		.proc_handler	= proc_dointvec_minmax,
 831		.extra1		= &pid_max_min,
 832		.extra2		= &pid_max_max,
 833	},
 834	{
 835		.procname	= "panic_on_oops",
 836		.data		= &panic_on_oops,
 837		.maxlen		= sizeof(int),
 838		.mode		= 0644,
 839		.proc_handler	= proc_dointvec,
 840	},
 841	{
 842		.procname	= "panic_print",
 843		.data		= &panic_print,
 844		.maxlen		= sizeof(unsigned long),
 845		.mode		= 0644,
 846		.proc_handler	= proc_doulongvec_minmax,
 847	},
 848#if defined CONFIG_PRINTK
 849	{
 850		.procname	= "printk",
 851		.data		= &console_loglevel,
 852		.maxlen		= 4*sizeof(int),
 853		.mode		= 0644,
 854		.proc_handler	= proc_dointvec,
 855	},
 856	{
 857		.procname	= "printk_ratelimit",
 858		.data		= &printk_ratelimit_state.interval,
 859		.maxlen		= sizeof(int),
 860		.mode		= 0644,
 861		.proc_handler	= proc_dointvec_jiffies,
 862	},
 863	{
 864		.procname	= "printk_ratelimit_burst",
 865		.data		= &printk_ratelimit_state.burst,
 866		.maxlen		= sizeof(int),
 867		.mode		= 0644,
 868		.proc_handler	= proc_dointvec,
 869	},
 870	{
 871		.procname	= "printk_delay",
 872		.data		= &printk_delay_msec,
 873		.maxlen		= sizeof(int),
 874		.mode		= 0644,
 875		.proc_handler	= proc_dointvec_minmax,
 876		.extra1		= SYSCTL_ZERO,
 877		.extra2		= &ten_thousand,
 878	},
 879	{
 880		.procname	= "printk_devkmsg",
 881		.data		= devkmsg_log_str,
 882		.maxlen		= DEVKMSG_STR_MAX_SIZE,
 883		.mode		= 0644,
 884		.proc_handler	= devkmsg_sysctl_set_loglvl,
 885	},
 886	{
 887		.procname	= "dmesg_restrict",
 888		.data		= &dmesg_restrict,
 889		.maxlen		= sizeof(int),
 890		.mode		= 0644,
 891		.proc_handler	= proc_dointvec_minmax_sysadmin,
 892		.extra1		= SYSCTL_ZERO,
 893		.extra2		= SYSCTL_ONE,
 894	},
 895	{
 896		.procname	= "kptr_restrict",
 897		.data		= &kptr_restrict,
 898		.maxlen		= sizeof(int),
 899		.mode		= 0644,
 900		.proc_handler	= proc_dointvec_minmax_sysadmin,
 901		.extra1		= SYSCTL_ZERO,
 902		.extra2		= &two,
 903	},
 904#endif
 905	{
 906		.procname	= "ngroups_max",
 907		.data		= &ngroups_max,
 908		.maxlen		= sizeof (int),
 909		.mode		= 0444,
 910		.proc_handler	= proc_dointvec,
 911	},
 912	{
 913		.procname	= "cap_last_cap",
 914		.data		= (void *)&cap_last_cap,
 915		.maxlen		= sizeof(int),
 916		.mode		= 0444,
 917		.proc_handler	= proc_dointvec,
 918	},
 919#if defined(CONFIG_LOCKUP_DETECTOR)
 920	{
 921		.procname       = "watchdog",
 922		.data		= &watchdog_user_enabled,
 923		.maxlen		= sizeof(int),
 924		.mode		= 0644,
 925		.proc_handler   = proc_watchdog,
 926		.extra1		= SYSCTL_ZERO,
 927		.extra2		= SYSCTL_ONE,
 928	},
 929	{
 930		.procname	= "watchdog_thresh",
 931		.data		= &watchdog_thresh,
 932		.maxlen		= sizeof(int),
 933		.mode		= 0644,
 934		.proc_handler	= proc_watchdog_thresh,
 935		.extra1		= SYSCTL_ZERO,
 936		.extra2		= &sixty,
 937	},
 938	{
 939		.procname       = "nmi_watchdog",
 940		.data		= &nmi_watchdog_user_enabled,
 941		.maxlen		= sizeof(int),
 942		.mode		= NMI_WATCHDOG_SYSCTL_PERM,
 943		.proc_handler   = proc_nmi_watchdog,
 944		.extra1		= SYSCTL_ZERO,
 945		.extra2		= SYSCTL_ONE,
 946	},
 947	{
 948		.procname	= "watchdog_cpumask",
 949		.data		= &watchdog_cpumask_bits,
 950		.maxlen		= NR_CPUS,
 951		.mode		= 0644,
 952		.proc_handler	= proc_watchdog_cpumask,
 953	},
 954#ifdef CONFIG_SOFTLOCKUP_DETECTOR
 955	{
 956		.procname       = "soft_watchdog",
 957		.data		= &soft_watchdog_user_enabled,
 958		.maxlen		= sizeof(int),
 959		.mode		= 0644,
 960		.proc_handler   = proc_soft_watchdog,
 961		.extra1		= SYSCTL_ZERO,
 962		.extra2		= SYSCTL_ONE,
 963	},
 964	{
 965		.procname	= "softlockup_panic",
 966		.data		= &softlockup_panic,
 967		.maxlen		= sizeof(int),
 968		.mode		= 0644,
 969		.proc_handler	= proc_dointvec_minmax,
 970		.extra1		= SYSCTL_ZERO,
 971		.extra2		= SYSCTL_ONE,
 972	},
 973#ifdef CONFIG_SMP
 974	{
 975		.procname	= "softlockup_all_cpu_backtrace",
 976		.data		= &sysctl_softlockup_all_cpu_backtrace,
 977		.maxlen		= sizeof(int),
 978		.mode		= 0644,
 979		.proc_handler	= proc_dointvec_minmax,
 980		.extra1		= SYSCTL_ZERO,
 981		.extra2		= SYSCTL_ONE,
 982	},
 983#endif /* CONFIG_SMP */
 984#endif
 985#ifdef CONFIG_HARDLOCKUP_DETECTOR
 986	{
 987		.procname	= "hardlockup_panic",
 988		.data		= &hardlockup_panic,
 989		.maxlen		= sizeof(int),
 990		.mode		= 0644,
 991		.proc_handler	= proc_dointvec_minmax,
 992		.extra1		= SYSCTL_ZERO,
 993		.extra2		= SYSCTL_ONE,
 994	},
 995#ifdef CONFIG_SMP
 996	{
 997		.procname	= "hardlockup_all_cpu_backtrace",
 998		.data		= &sysctl_hardlockup_all_cpu_backtrace,
 999		.maxlen		= sizeof(int),
1000		.mode		= 0644,
1001		.proc_handler	= proc_dointvec_minmax,
1002		.extra1		= SYSCTL_ZERO,
1003		.extra2		= SYSCTL_ONE,
1004	},
1005#endif /* CONFIG_SMP */
1006#endif
1007#endif
1008
1009#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
1010	{
1011		.procname       = "unknown_nmi_panic",
1012		.data           = &unknown_nmi_panic,
1013		.maxlen         = sizeof (int),
1014		.mode           = 0644,
1015		.proc_handler   = proc_dointvec,
1016	},
1017#endif
1018#if defined(CONFIG_X86)
 
 
1019	{
1020		.procname	= "panic_on_unrecovered_nmi",
1021		.data		= &panic_on_unrecovered_nmi,
1022		.maxlen		= sizeof(int),
1023		.mode		= 0644,
1024		.proc_handler	= proc_dointvec,
1025	},
 
 
1026	{
1027		.procname	= "panic_on_io_nmi",
1028		.data		= &panic_on_io_nmi,
1029		.maxlen		= sizeof(int),
1030		.mode		= 0644,
1031		.proc_handler	= proc_dointvec,
1032	},
1033#ifdef CONFIG_DEBUG_STACKOVERFLOW
1034	{
1035		.procname	= "panic_on_stackoverflow",
1036		.data		= &sysctl_panic_on_stackoverflow,
1037		.maxlen		= sizeof(int),
1038		.mode		= 0644,
1039		.proc_handler	= proc_dointvec,
1040	},
1041#endif
1042	{
1043		.procname	= "bootloader_type",
1044		.data		= &bootloader_type,
1045		.maxlen		= sizeof (int),
1046		.mode		= 0444,
1047		.proc_handler	= proc_dointvec,
1048	},
1049	{
1050		.procname	= "bootloader_version",
1051		.data		= &bootloader_version,
1052		.maxlen		= sizeof (int),
1053		.mode		= 0444,
1054		.proc_handler	= proc_dointvec,
1055	},
1056	{
1057		.procname	= "io_delay_type",
1058		.data		= &io_delay_type,
1059		.maxlen		= sizeof(int),
1060		.mode		= 0644,
1061		.proc_handler	= proc_dointvec,
1062	},
1063#endif
1064#if defined(CONFIG_MMU)
1065	{
1066		.procname	= "randomize_va_space",
1067		.data		= &randomize_va_space,
1068		.maxlen		= sizeof(int),
1069		.mode		= 0644,
1070		.proc_handler	= proc_dointvec,
1071	},
1072#endif
1073#if defined(CONFIG_S390) && defined(CONFIG_SMP)
1074	{
1075		.procname	= "spin_retry",
1076		.data		= &spin_retry,
1077		.maxlen		= sizeof (int),
1078		.mode		= 0644,
1079		.proc_handler	= proc_dointvec,
1080	},
1081#endif
1082#if	defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
1083	{
1084		.procname	= "acpi_video_flags",
1085		.data		= &acpi_realmode_flags,
1086		.maxlen		= sizeof (unsigned long),
1087		.mode		= 0644,
1088		.proc_handler	= proc_doulongvec_minmax,
1089	},
1090#endif
1091#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
1092	{
1093		.procname	= "ignore-unaligned-usertrap",
1094		.data		= &no_unaligned_warning,
1095		.maxlen		= sizeof (int),
1096	 	.mode		= 0644,
1097		.proc_handler	= proc_dointvec,
1098	},
1099#endif
1100#ifdef CONFIG_IA64
1101	{
1102		.procname	= "unaligned-dump-stack",
1103		.data		= &unaligned_dump_stack,
1104		.maxlen		= sizeof (int),
1105		.mode		= 0644,
1106		.proc_handler	= proc_dointvec,
1107	},
1108#endif
1109#ifdef CONFIG_DETECT_HUNG_TASK
 
 
 
 
 
 
 
 
 
 
 
1110	{
1111		.procname	= "hung_task_panic",
1112		.data		= &sysctl_hung_task_panic,
1113		.maxlen		= sizeof(int),
1114		.mode		= 0644,
1115		.proc_handler	= proc_dointvec_minmax,
1116		.extra1		= SYSCTL_ZERO,
1117		.extra2		= SYSCTL_ONE,
1118	},
1119	{
1120		.procname	= "hung_task_check_count",
1121		.data		= &sysctl_hung_task_check_count,
1122		.maxlen		= sizeof(int),
1123		.mode		= 0644,
1124		.proc_handler	= proc_dointvec_minmax,
1125		.extra1		= SYSCTL_ZERO,
1126	},
1127	{
1128		.procname	= "hung_task_timeout_secs",
1129		.data		= &sysctl_hung_task_timeout_secs,
1130		.maxlen		= sizeof(unsigned long),
1131		.mode		= 0644,
1132		.proc_handler	= proc_dohung_task_timeout_secs,
1133		.extra2		= &hung_task_timeout_max,
1134	},
1135	{
1136		.procname	= "hung_task_check_interval_secs",
1137		.data		= &sysctl_hung_task_check_interval_secs,
1138		.maxlen		= sizeof(unsigned long),
1139		.mode		= 0644,
1140		.proc_handler	= proc_dohung_task_timeout_secs,
1141		.extra2		= &hung_task_timeout_max,
1142	},
1143	{
1144		.procname	= "hung_task_warnings",
1145		.data		= &sysctl_hung_task_warnings,
1146		.maxlen		= sizeof(int),
1147		.mode		= 0644,
1148		.proc_handler	= proc_dointvec_minmax,
1149		.extra1		= &neg_one,
1150	},
1151#endif
1152#ifdef CONFIG_RT_MUTEXES
1153	{
1154		.procname	= "max_lock_depth",
1155		.data		= &max_lock_depth,
1156		.maxlen		= sizeof(int),
1157		.mode		= 0644,
1158		.proc_handler	= proc_dointvec,
1159	},
1160#endif
1161	{
1162		.procname	= "poweroff_cmd",
1163		.data		= &poweroff_cmd,
1164		.maxlen		= POWEROFF_CMD_PATH_LEN,
1165		.mode		= 0644,
1166		.proc_handler	= proc_dostring,
1167	},
1168#ifdef CONFIG_KEYS
1169	{
1170		.procname	= "keys",
1171		.mode		= 0555,
1172		.child		= key_sysctls,
1173	},
1174#endif
1175#ifdef CONFIG_PERF_EVENTS
1176	/*
1177	 * User-space scripts rely on the existence of this file
1178	 * as a feature check for perf_events being enabled.
1179	 *
1180	 * So it's an ABI, do not remove!
1181	 */
1182	{
1183		.procname	= "perf_event_paranoid",
1184		.data		= &sysctl_perf_event_paranoid,
1185		.maxlen		= sizeof(sysctl_perf_event_paranoid),
1186		.mode		= 0644,
1187		.proc_handler	= proc_dointvec,
1188	},
1189	{
1190		.procname	= "perf_event_mlock_kb",
1191		.data		= &sysctl_perf_event_mlock,
1192		.maxlen		= sizeof(sysctl_perf_event_mlock),
1193		.mode		= 0644,
1194		.proc_handler	= proc_dointvec,
1195	},
1196	{
1197		.procname	= "perf_event_max_sample_rate",
1198		.data		= &sysctl_perf_event_sample_rate,
1199		.maxlen		= sizeof(sysctl_perf_event_sample_rate),
1200		.mode		= 0644,
1201		.proc_handler	= perf_proc_update_handler,
1202		.extra1		= SYSCTL_ONE,
1203	},
1204	{
1205		.procname	= "perf_cpu_time_max_percent",
1206		.data		= &sysctl_perf_cpu_time_max_percent,
1207		.maxlen		= sizeof(sysctl_perf_cpu_time_max_percent),
1208		.mode		= 0644,
1209		.proc_handler	= perf_cpu_time_max_percent_handler,
1210		.extra1		= SYSCTL_ZERO,
1211		.extra2		= &one_hundred,
1212	},
1213	{
1214		.procname	= "perf_event_max_stack",
1215		.data		= &sysctl_perf_event_max_stack,
1216		.maxlen		= sizeof(sysctl_perf_event_max_stack),
1217		.mode		= 0644,
1218		.proc_handler	= perf_event_max_stack_handler,
1219		.extra1		= SYSCTL_ZERO,
1220		.extra2		= &six_hundred_forty_kb,
1221	},
1222	{
1223		.procname	= "perf_event_max_contexts_per_stack",
1224		.data		= &sysctl_perf_event_max_contexts_per_stack,
1225		.maxlen		= sizeof(sysctl_perf_event_max_contexts_per_stack),
1226		.mode		= 0644,
1227		.proc_handler	= perf_event_max_stack_handler,
1228		.extra1		= SYSCTL_ZERO,
1229		.extra2		= &one_thousand,
1230	},
1231#endif
1232	{
1233		.procname	= "panic_on_warn",
1234		.data		= &panic_on_warn,
1235		.maxlen		= sizeof(int),
1236		.mode		= 0644,
1237		.proc_handler	= proc_dointvec_minmax,
1238		.extra1		= SYSCTL_ZERO,
1239		.extra2		= SYSCTL_ONE,
1240	},
1241#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
1242	{
1243		.procname	= "timer_migration",
1244		.data		= &sysctl_timer_migration,
1245		.maxlen		= sizeof(unsigned int),
1246		.mode		= 0644,
1247		.proc_handler	= timer_migration_handler,
1248		.extra1		= SYSCTL_ZERO,
1249		.extra2		= SYSCTL_ONE,
1250	},
1251#endif
1252#ifdef CONFIG_BPF_SYSCALL
1253	{
1254		.procname	= "unprivileged_bpf_disabled",
1255		.data		= &sysctl_unprivileged_bpf_disabled,
1256		.maxlen		= sizeof(sysctl_unprivileged_bpf_disabled),
1257		.mode		= 0644,
1258		/* only handle a transition from default "0" to "1" */
1259		.proc_handler	= proc_dointvec_minmax,
1260		.extra1		= SYSCTL_ONE,
1261		.extra2		= SYSCTL_ONE,
1262	},
1263	{
1264		.procname	= "bpf_stats_enabled",
1265		.data		= &bpf_stats_enabled_key.key,
1266		.maxlen		= sizeof(bpf_stats_enabled_key),
1267		.mode		= 0644,
1268		.proc_handler	= proc_do_static_key,
1269	},
1270#endif
1271#if defined(CONFIG_TREE_RCU) || defined(CONFIG_PREEMPT_RCU)
1272	{
1273		.procname	= "panic_on_rcu_stall",
1274		.data		= &sysctl_panic_on_rcu_stall,
1275		.maxlen		= sizeof(sysctl_panic_on_rcu_stall),
1276		.mode		= 0644,
1277		.proc_handler	= proc_dointvec_minmax,
1278		.extra1		= SYSCTL_ZERO,
1279		.extra2		= SYSCTL_ONE,
1280	},
1281#endif
1282#ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
1283	{
1284		.procname	= "stack_erasing",
1285		.data		= NULL,
1286		.maxlen		= sizeof(int),
1287		.mode		= 0600,
1288		.proc_handler	= stack_erasing_sysctl,
1289		.extra1		= SYSCTL_ZERO,
1290		.extra2		= SYSCTL_ONE,
1291	},
1292#endif
1293	{ }
1294};
1295
1296static struct ctl_table vm_table[] = {
1297	{
1298		.procname	= "overcommit_memory",
1299		.data		= &sysctl_overcommit_memory,
1300		.maxlen		= sizeof(sysctl_overcommit_memory),
1301		.mode		= 0644,
1302		.proc_handler	= proc_dointvec_minmax,
1303		.extra1		= SYSCTL_ZERO,
1304		.extra2		= &two,
1305	},
1306	{
1307		.procname	= "panic_on_oom",
1308		.data		= &sysctl_panic_on_oom,
1309		.maxlen		= sizeof(sysctl_panic_on_oom),
1310		.mode		= 0644,
1311		.proc_handler	= proc_dointvec_minmax,
1312		.extra1		= SYSCTL_ZERO,
1313		.extra2		= &two,
1314	},
1315	{
1316		.procname	= "oom_kill_allocating_task",
1317		.data		= &sysctl_oom_kill_allocating_task,
1318		.maxlen		= sizeof(sysctl_oom_kill_allocating_task),
1319		.mode		= 0644,
1320		.proc_handler	= proc_dointvec,
1321	},
1322	{
1323		.procname	= "oom_dump_tasks",
1324		.data		= &sysctl_oom_dump_tasks,
1325		.maxlen		= sizeof(sysctl_oom_dump_tasks),
1326		.mode		= 0644,
1327		.proc_handler	= proc_dointvec,
1328	},
1329	{
1330		.procname	= "overcommit_ratio",
1331		.data		= &sysctl_overcommit_ratio,
1332		.maxlen		= sizeof(sysctl_overcommit_ratio),
1333		.mode		= 0644,
1334		.proc_handler	= overcommit_ratio_handler,
1335	},
1336	{
1337		.procname	= "overcommit_kbytes",
1338		.data		= &sysctl_overcommit_kbytes,
1339		.maxlen		= sizeof(sysctl_overcommit_kbytes),
1340		.mode		= 0644,
1341		.proc_handler	= overcommit_kbytes_handler,
1342	},
1343	{
1344		.procname	= "page-cluster", 
1345		.data		= &page_cluster,
1346		.maxlen		= sizeof(int),
1347		.mode		= 0644,
1348		.proc_handler	= proc_dointvec_minmax,
1349		.extra1		= SYSCTL_ZERO,
1350	},
1351	{
1352		.procname	= "dirty_background_ratio",
1353		.data		= &dirty_background_ratio,
1354		.maxlen		= sizeof(dirty_background_ratio),
1355		.mode		= 0644,
1356		.proc_handler	= dirty_background_ratio_handler,
1357		.extra1		= SYSCTL_ZERO,
1358		.extra2		= &one_hundred,
1359	},
1360	{
1361		.procname	= "dirty_background_bytes",
1362		.data		= &dirty_background_bytes,
1363		.maxlen		= sizeof(dirty_background_bytes),
1364		.mode		= 0644,
1365		.proc_handler	= dirty_background_bytes_handler,
1366		.extra1		= &one_ul,
1367	},
1368	{
1369		.procname	= "dirty_ratio",
1370		.data		= &vm_dirty_ratio,
1371		.maxlen		= sizeof(vm_dirty_ratio),
1372		.mode		= 0644,
1373		.proc_handler	= dirty_ratio_handler,
1374		.extra1		= SYSCTL_ZERO,
1375		.extra2		= &one_hundred,
1376	},
1377	{
1378		.procname	= "dirty_bytes",
1379		.data		= &vm_dirty_bytes,
1380		.maxlen		= sizeof(vm_dirty_bytes),
1381		.mode		= 0644,
1382		.proc_handler	= dirty_bytes_handler,
1383		.extra1		= &dirty_bytes_min,
1384	},
1385	{
1386		.procname	= "dirty_writeback_centisecs",
1387		.data		= &dirty_writeback_interval,
1388		.maxlen		= sizeof(dirty_writeback_interval),
1389		.mode		= 0644,
1390		.proc_handler	= dirty_writeback_centisecs_handler,
1391	},
1392	{
1393		.procname	= "dirty_expire_centisecs",
1394		.data		= &dirty_expire_interval,
1395		.maxlen		= sizeof(dirty_expire_interval),
1396		.mode		= 0644,
1397		.proc_handler	= proc_dointvec_minmax,
1398		.extra1		= SYSCTL_ZERO,
1399	},
1400	{
1401		.procname	= "dirtytime_expire_seconds",
1402		.data		= &dirtytime_expire_interval,
1403		.maxlen		= sizeof(dirtytime_expire_interval),
1404		.mode		= 0644,
1405		.proc_handler	= dirtytime_interval_handler,
1406		.extra1		= SYSCTL_ZERO,
1407	},
1408	{
1409		.procname	= "swappiness",
1410		.data		= &vm_swappiness,
1411		.maxlen		= sizeof(vm_swappiness),
1412		.mode		= 0644,
1413		.proc_handler	= proc_dointvec_minmax,
1414		.extra1		= SYSCTL_ZERO,
1415		.extra2		= &one_hundred,
1416	},
1417#ifdef CONFIG_HUGETLB_PAGE
1418	{
1419		.procname	= "nr_hugepages",
1420		.data		= NULL,
1421		.maxlen		= sizeof(unsigned long),
1422		.mode		= 0644,
1423		.proc_handler	= hugetlb_sysctl_handler,
1424	},
1425#ifdef CONFIG_NUMA
1426	{
1427		.procname       = "nr_hugepages_mempolicy",
1428		.data           = NULL,
1429		.maxlen         = sizeof(unsigned long),
1430		.mode           = 0644,
1431		.proc_handler   = &hugetlb_mempolicy_sysctl_handler,
1432	},
1433	{
1434		.procname		= "numa_stat",
1435		.data			= &sysctl_vm_numa_stat,
1436		.maxlen			= sizeof(int),
1437		.mode			= 0644,
1438		.proc_handler	= sysctl_vm_numa_stat_handler,
1439		.extra1			= SYSCTL_ZERO,
1440		.extra2			= SYSCTL_ONE,
1441	},
1442#endif
1443	 {
1444		.procname	= "hugetlb_shm_group",
1445		.data		= &sysctl_hugetlb_shm_group,
1446		.maxlen		= sizeof(gid_t),
1447		.mode		= 0644,
1448		.proc_handler	= proc_dointvec,
1449	 },
1450	{
1451		.procname	= "nr_overcommit_hugepages",
1452		.data		= NULL,
1453		.maxlen		= sizeof(unsigned long),
1454		.mode		= 0644,
1455		.proc_handler	= hugetlb_overcommit_handler,
1456	},
1457#endif
1458	{
1459		.procname	= "lowmem_reserve_ratio",
1460		.data		= &sysctl_lowmem_reserve_ratio,
1461		.maxlen		= sizeof(sysctl_lowmem_reserve_ratio),
1462		.mode		= 0644,
1463		.proc_handler	= lowmem_reserve_ratio_sysctl_handler,
1464	},
1465	{
1466		.procname	= "drop_caches",
1467		.data		= &sysctl_drop_caches,
1468		.maxlen		= sizeof(int),
1469		.mode		= 0644,
1470		.proc_handler	= drop_caches_sysctl_handler,
1471		.extra1		= SYSCTL_ONE,
1472		.extra2		= &four,
1473	},
1474#ifdef CONFIG_COMPACTION
1475	{
1476		.procname	= "compact_memory",
1477		.data		= &sysctl_compact_memory,
1478		.maxlen		= sizeof(int),
1479		.mode		= 0200,
1480		.proc_handler	= sysctl_compaction_handler,
1481	},
1482	{
 
 
 
 
 
 
 
 
 
1483		.procname	= "extfrag_threshold",
1484		.data		= &sysctl_extfrag_threshold,
1485		.maxlen		= sizeof(int),
1486		.mode		= 0644,
1487		.proc_handler	= proc_dointvec_minmax,
1488		.extra1		= &min_extfrag_threshold,
1489		.extra2		= &max_extfrag_threshold,
1490	},
1491	{
1492		.procname	= "compact_unevictable_allowed",
1493		.data		= &sysctl_compact_unevictable_allowed,
1494		.maxlen		= sizeof(int),
1495		.mode		= 0644,
1496		.proc_handler	= proc_dointvec,
1497		.extra1		= SYSCTL_ZERO,
1498		.extra2		= SYSCTL_ONE,
1499	},
1500
1501#endif /* CONFIG_COMPACTION */
1502	{
1503		.procname	= "min_free_kbytes",
1504		.data		= &min_free_kbytes,
1505		.maxlen		= sizeof(min_free_kbytes),
1506		.mode		= 0644,
1507		.proc_handler	= min_free_kbytes_sysctl_handler,
1508		.extra1		= SYSCTL_ZERO,
1509	},
1510	{
1511		.procname	= "watermark_boost_factor",
1512		.data		= &watermark_boost_factor,
1513		.maxlen		= sizeof(watermark_boost_factor),
1514		.mode		= 0644,
1515		.proc_handler	= watermark_boost_factor_sysctl_handler,
1516		.extra1		= SYSCTL_ZERO,
1517	},
1518	{
1519		.procname	= "watermark_scale_factor",
1520		.data		= &watermark_scale_factor,
1521		.maxlen		= sizeof(watermark_scale_factor),
1522		.mode		= 0644,
1523		.proc_handler	= watermark_scale_factor_sysctl_handler,
1524		.extra1		= SYSCTL_ONE,
1525		.extra2		= &one_thousand,
1526	},
1527	{
1528		.procname	= "percpu_pagelist_fraction",
1529		.data		= &percpu_pagelist_fraction,
1530		.maxlen		= sizeof(percpu_pagelist_fraction),
1531		.mode		= 0644,
1532		.proc_handler	= percpu_pagelist_fraction_sysctl_handler,
1533		.extra1		= SYSCTL_ZERO,
1534	},
 
 
 
 
 
 
 
 
1535#ifdef CONFIG_MMU
1536	{
1537		.procname	= "max_map_count",
1538		.data		= &sysctl_max_map_count,
1539		.maxlen		= sizeof(sysctl_max_map_count),
1540		.mode		= 0644,
1541		.proc_handler	= proc_dointvec_minmax,
1542		.extra1		= SYSCTL_ZERO,
1543	},
1544#else
1545	{
1546		.procname	= "nr_trim_pages",
1547		.data		= &sysctl_nr_trim_pages,
1548		.maxlen		= sizeof(sysctl_nr_trim_pages),
1549		.mode		= 0644,
1550		.proc_handler	= proc_dointvec_minmax,
1551		.extra1		= SYSCTL_ZERO,
1552	},
1553#endif
1554	{
1555		.procname	= "laptop_mode",
1556		.data		= &laptop_mode,
1557		.maxlen		= sizeof(laptop_mode),
1558		.mode		= 0644,
1559		.proc_handler	= proc_dointvec_jiffies,
1560	},
1561	{
1562		.procname	= "block_dump",
1563		.data		= &block_dump,
1564		.maxlen		= sizeof(block_dump),
1565		.mode		= 0644,
1566		.proc_handler	= proc_dointvec,
1567		.extra1		= SYSCTL_ZERO,
1568	},
1569	{
1570		.procname	= "vfs_cache_pressure",
1571		.data		= &sysctl_vfs_cache_pressure,
1572		.maxlen		= sizeof(sysctl_vfs_cache_pressure),
1573		.mode		= 0644,
1574		.proc_handler	= proc_dointvec,
1575		.extra1		= SYSCTL_ZERO,
1576	},
1577#if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
1578    defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
1579	{
1580		.procname	= "legacy_va_layout",
1581		.data		= &sysctl_legacy_va_layout,
1582		.maxlen		= sizeof(sysctl_legacy_va_layout),
1583		.mode		= 0644,
1584		.proc_handler	= proc_dointvec,
1585		.extra1		= SYSCTL_ZERO,
1586	},
1587#endif
1588#ifdef CONFIG_NUMA
1589	{
1590		.procname	= "zone_reclaim_mode",
1591		.data		= &node_reclaim_mode,
1592		.maxlen		= sizeof(node_reclaim_mode),
1593		.mode		= 0644,
1594		.proc_handler	= proc_dointvec,
1595		.extra1		= SYSCTL_ZERO,
1596	},
1597	{
1598		.procname	= "min_unmapped_ratio",
1599		.data		= &sysctl_min_unmapped_ratio,
1600		.maxlen		= sizeof(sysctl_min_unmapped_ratio),
1601		.mode		= 0644,
1602		.proc_handler	= sysctl_min_unmapped_ratio_sysctl_handler,
1603		.extra1		= SYSCTL_ZERO,
1604		.extra2		= &one_hundred,
1605	},
1606	{
1607		.procname	= "min_slab_ratio",
1608		.data		= &sysctl_min_slab_ratio,
1609		.maxlen		= sizeof(sysctl_min_slab_ratio),
1610		.mode		= 0644,
1611		.proc_handler	= sysctl_min_slab_ratio_sysctl_handler,
1612		.extra1		= SYSCTL_ZERO,
1613		.extra2		= &one_hundred,
1614	},
1615#endif
1616#ifdef CONFIG_SMP
1617	{
1618		.procname	= "stat_interval",
1619		.data		= &sysctl_stat_interval,
1620		.maxlen		= sizeof(sysctl_stat_interval),
1621		.mode		= 0644,
1622		.proc_handler	= proc_dointvec_jiffies,
1623	},
1624	{
1625		.procname	= "stat_refresh",
1626		.data		= NULL,
1627		.maxlen		= 0,
1628		.mode		= 0600,
1629		.proc_handler	= vmstat_refresh,
1630	},
1631#endif
1632#ifdef CONFIG_MMU
1633	{
1634		.procname	= "mmap_min_addr",
1635		.data		= &dac_mmap_min_addr,
1636		.maxlen		= sizeof(unsigned long),
1637		.mode		= 0644,
1638		.proc_handler	= mmap_min_addr_handler,
1639	},
1640#endif
1641#ifdef CONFIG_NUMA
1642	{
1643		.procname	= "numa_zonelist_order",
1644		.data		= &numa_zonelist_order,
1645		.maxlen		= NUMA_ZONELIST_ORDER_LEN,
1646		.mode		= 0644,
1647		.proc_handler	= numa_zonelist_order_handler,
1648	},
1649#endif
1650#if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
1651   (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
1652	{
1653		.procname	= "vdso_enabled",
1654#ifdef CONFIG_X86_32
1655		.data		= &vdso32_enabled,
1656		.maxlen		= sizeof(vdso32_enabled),
1657#else
1658		.data		= &vdso_enabled,
1659		.maxlen		= sizeof(vdso_enabled),
1660#endif
1661		.mode		= 0644,
1662		.proc_handler	= proc_dointvec,
1663		.extra1		= SYSCTL_ZERO,
1664	},
1665#endif
1666#ifdef CONFIG_HIGHMEM
1667	{
1668		.procname	= "highmem_is_dirtyable",
1669		.data		= &vm_highmem_is_dirtyable,
1670		.maxlen		= sizeof(vm_highmem_is_dirtyable),
1671		.mode		= 0644,
1672		.proc_handler	= proc_dointvec_minmax,
1673		.extra1		= SYSCTL_ZERO,
1674		.extra2		= SYSCTL_ONE,
1675	},
1676#endif
1677#ifdef CONFIG_MEMORY_FAILURE
1678	{
1679		.procname	= "memory_failure_early_kill",
1680		.data		= &sysctl_memory_failure_early_kill,
1681		.maxlen		= sizeof(sysctl_memory_failure_early_kill),
1682		.mode		= 0644,
1683		.proc_handler	= proc_dointvec_minmax,
1684		.extra1		= SYSCTL_ZERO,
1685		.extra2		= SYSCTL_ONE,
1686	},
1687	{
1688		.procname	= "memory_failure_recovery",
1689		.data		= &sysctl_memory_failure_recovery,
1690		.maxlen		= sizeof(sysctl_memory_failure_recovery),
1691		.mode		= 0644,
1692		.proc_handler	= proc_dointvec_minmax,
1693		.extra1		= SYSCTL_ZERO,
1694		.extra2		= SYSCTL_ONE,
1695	},
1696#endif
1697	{
1698		.procname	= "user_reserve_kbytes",
1699		.data		= &sysctl_user_reserve_kbytes,
1700		.maxlen		= sizeof(sysctl_user_reserve_kbytes),
1701		.mode		= 0644,
1702		.proc_handler	= proc_doulongvec_minmax,
1703	},
1704	{
1705		.procname	= "admin_reserve_kbytes",
1706		.data		= &sysctl_admin_reserve_kbytes,
1707		.maxlen		= sizeof(sysctl_admin_reserve_kbytes),
1708		.mode		= 0644,
1709		.proc_handler	= proc_doulongvec_minmax,
1710	},
1711#ifdef CONFIG_HAVE_ARCH_MMAP_RND_BITS
1712	{
1713		.procname	= "mmap_rnd_bits",
1714		.data		= &mmap_rnd_bits,
1715		.maxlen		= sizeof(mmap_rnd_bits),
1716		.mode		= 0600,
1717		.proc_handler	= proc_dointvec_minmax,
1718		.extra1		= (void *)&mmap_rnd_bits_min,
1719		.extra2		= (void *)&mmap_rnd_bits_max,
1720	},
1721#endif
1722#ifdef CONFIG_HAVE_ARCH_MMAP_RND_COMPAT_BITS
1723	{
1724		.procname	= "mmap_rnd_compat_bits",
1725		.data		= &mmap_rnd_compat_bits,
1726		.maxlen		= sizeof(mmap_rnd_compat_bits),
1727		.mode		= 0600,
1728		.proc_handler	= proc_dointvec_minmax,
1729		.extra1		= (void *)&mmap_rnd_compat_bits_min,
1730		.extra2		= (void *)&mmap_rnd_compat_bits_max,
1731	},
1732#endif
1733#ifdef CONFIG_USERFAULTFD
1734	{
1735		.procname	= "unprivileged_userfaultfd",
1736		.data		= &sysctl_unprivileged_userfaultfd,
1737		.maxlen		= sizeof(sysctl_unprivileged_userfaultfd),
1738		.mode		= 0644,
1739		.proc_handler	= proc_dointvec_minmax,
1740		.extra1		= SYSCTL_ZERO,
1741		.extra2		= SYSCTL_ONE,
1742	},
1743#endif
1744	{ }
1745};
1746
1747static struct ctl_table fs_table[] = {
1748	{
1749		.procname	= "inode-nr",
1750		.data		= &inodes_stat,
1751		.maxlen		= 2*sizeof(long),
1752		.mode		= 0444,
1753		.proc_handler	= proc_nr_inodes,
1754	},
1755	{
1756		.procname	= "inode-state",
1757		.data		= &inodes_stat,
1758		.maxlen		= 7*sizeof(long),
1759		.mode		= 0444,
1760		.proc_handler	= proc_nr_inodes,
1761	},
1762	{
1763		.procname	= "file-nr",
1764		.data		= &files_stat,
1765		.maxlen		= sizeof(files_stat),
1766		.mode		= 0444,
1767		.proc_handler	= proc_nr_files,
1768	},
1769	{
1770		.procname	= "file-max",
1771		.data		= &files_stat.max_files,
1772		.maxlen		= sizeof(files_stat.max_files),
1773		.mode		= 0644,
1774		.proc_handler	= proc_doulongvec_minmax,
1775		.extra1		= &zero_ul,
1776		.extra2		= &long_max,
1777	},
1778	{
1779		.procname	= "nr_open",
1780		.data		= &sysctl_nr_open,
1781		.maxlen		= sizeof(unsigned int),
1782		.mode		= 0644,
1783		.proc_handler	= proc_dointvec_minmax,
1784		.extra1		= &sysctl_nr_open_min,
1785		.extra2		= &sysctl_nr_open_max,
1786	},
1787	{
1788		.procname	= "dentry-state",
1789		.data		= &dentry_stat,
1790		.maxlen		= 6*sizeof(long),
1791		.mode		= 0444,
1792		.proc_handler	= proc_nr_dentry,
1793	},
1794	{
1795		.procname	= "overflowuid",
1796		.data		= &fs_overflowuid,
1797		.maxlen		= sizeof(int),
1798		.mode		= 0644,
1799		.proc_handler	= proc_dointvec_minmax,
1800		.extra1		= &minolduid,
1801		.extra2		= &maxolduid,
1802	},
1803	{
1804		.procname	= "overflowgid",
1805		.data		= &fs_overflowgid,
1806		.maxlen		= sizeof(int),
1807		.mode		= 0644,
1808		.proc_handler	= proc_dointvec_minmax,
1809		.extra1		= &minolduid,
1810		.extra2		= &maxolduid,
1811	},
1812#ifdef CONFIG_FILE_LOCKING
1813	{
1814		.procname	= "leases-enable",
1815		.data		= &leases_enable,
1816		.maxlen		= sizeof(int),
1817		.mode		= 0644,
1818		.proc_handler	= proc_dointvec,
1819	},
1820#endif
1821#ifdef CONFIG_DNOTIFY
1822	{
1823		.procname	= "dir-notify-enable",
1824		.data		= &dir_notify_enable,
1825		.maxlen		= sizeof(int),
1826		.mode		= 0644,
1827		.proc_handler	= proc_dointvec,
1828	},
1829#endif
1830#ifdef CONFIG_MMU
1831#ifdef CONFIG_FILE_LOCKING
1832	{
1833		.procname	= "lease-break-time",
1834		.data		= &lease_break_time,
1835		.maxlen		= sizeof(int),
1836		.mode		= 0644,
1837		.proc_handler	= proc_dointvec,
1838	},
1839#endif
1840#ifdef CONFIG_AIO
1841	{
1842		.procname	= "aio-nr",
1843		.data		= &aio_nr,
1844		.maxlen		= sizeof(aio_nr),
1845		.mode		= 0444,
1846		.proc_handler	= proc_doulongvec_minmax,
1847	},
1848	{
1849		.procname	= "aio-max-nr",
1850		.data		= &aio_max_nr,
1851		.maxlen		= sizeof(aio_max_nr),
1852		.mode		= 0644,
1853		.proc_handler	= proc_doulongvec_minmax,
1854	},
1855#endif /* CONFIG_AIO */
1856#ifdef CONFIG_INOTIFY_USER
1857	{
1858		.procname	= "inotify",
1859		.mode		= 0555,
1860		.child		= inotify_table,
1861	},
1862#endif	
1863#ifdef CONFIG_EPOLL
1864	{
1865		.procname	= "epoll",
1866		.mode		= 0555,
1867		.child		= epoll_table,
1868	},
1869#endif
1870#endif
1871	{
1872		.procname	= "protected_symlinks",
1873		.data		= &sysctl_protected_symlinks,
1874		.maxlen		= sizeof(int),
1875		.mode		= 0600,
1876		.proc_handler	= proc_dointvec_minmax,
1877		.extra1		= SYSCTL_ZERO,
1878		.extra2		= SYSCTL_ONE,
1879	},
1880	{
1881		.procname	= "protected_hardlinks",
1882		.data		= &sysctl_protected_hardlinks,
1883		.maxlen		= sizeof(int),
1884		.mode		= 0600,
1885		.proc_handler	= proc_dointvec_minmax,
1886		.extra1		= SYSCTL_ZERO,
1887		.extra2		= SYSCTL_ONE,
1888	},
1889	{
1890		.procname	= "protected_fifos",
1891		.data		= &sysctl_protected_fifos,
1892		.maxlen		= sizeof(int),
1893		.mode		= 0600,
1894		.proc_handler	= proc_dointvec_minmax,
1895		.extra1		= SYSCTL_ZERO,
1896		.extra2		= &two,
1897	},
1898	{
1899		.procname	= "protected_regular",
1900		.data		= &sysctl_protected_regular,
1901		.maxlen		= sizeof(int),
1902		.mode		= 0600,
1903		.proc_handler	= proc_dointvec_minmax,
1904		.extra1		= SYSCTL_ZERO,
1905		.extra2		= &two,
1906	},
1907	{
1908		.procname	= "suid_dumpable",
1909		.data		= &suid_dumpable,
1910		.maxlen		= sizeof(int),
1911		.mode		= 0644,
1912		.proc_handler	= proc_dointvec_minmax_coredump,
1913		.extra1		= SYSCTL_ZERO,
1914		.extra2		= &two,
1915	},
1916#if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1917	{
1918		.procname	= "binfmt_misc",
1919		.mode		= 0555,
1920		.child		= sysctl_mount_point,
1921	},
1922#endif
1923	{
1924		.procname	= "pipe-max-size",
1925		.data		= &pipe_max_size,
1926		.maxlen		= sizeof(pipe_max_size),
1927		.mode		= 0644,
1928		.proc_handler	= proc_dopipe_max_size,
1929	},
1930	{
1931		.procname	= "pipe-user-pages-hard",
1932		.data		= &pipe_user_pages_hard,
1933		.maxlen		= sizeof(pipe_user_pages_hard),
1934		.mode		= 0644,
1935		.proc_handler	= proc_doulongvec_minmax,
1936	},
1937	{
1938		.procname	= "pipe-user-pages-soft",
1939		.data		= &pipe_user_pages_soft,
1940		.maxlen		= sizeof(pipe_user_pages_soft),
1941		.mode		= 0644,
1942		.proc_handler	= proc_doulongvec_minmax,
1943	},
1944	{
1945		.procname	= "mount-max",
1946		.data		= &sysctl_mount_max,
1947		.maxlen		= sizeof(unsigned int),
1948		.mode		= 0644,
1949		.proc_handler	= proc_dointvec_minmax,
1950		.extra1		= SYSCTL_ONE,
1951	},
1952	{ }
1953};
1954
1955static struct ctl_table debug_table[] = {
1956#ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
1957	{
1958		.procname	= "exception-trace",
1959		.data		= &show_unhandled_signals,
1960		.maxlen		= sizeof(int),
1961		.mode		= 0644,
1962		.proc_handler	= proc_dointvec
1963	},
1964#endif
1965#if defined(CONFIG_OPTPROBES)
1966	{
1967		.procname	= "kprobes-optimization",
1968		.data		= &sysctl_kprobes_optimization,
1969		.maxlen		= sizeof(int),
1970		.mode		= 0644,
1971		.proc_handler	= proc_kprobes_optimization_handler,
1972		.extra1		= SYSCTL_ZERO,
1973		.extra2		= SYSCTL_ONE,
1974	},
1975#endif
1976	{ }
1977};
1978
1979static struct ctl_table dev_table[] = {
1980	{ }
1981};
1982
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1983int __init sysctl_init(void)
1984{
1985	struct ctl_table_header *hdr;
1986
1987	hdr = register_sysctl_table(sysctl_base_table);
1988	kmemleak_not_leak(hdr);
1989	return 0;
1990}
1991
1992#endif /* CONFIG_SYSCTL */
1993
1994/*
1995 * /proc/sys support
1996 */
1997
1998#ifdef CONFIG_PROC_SYSCTL
1999
2000static int _proc_do_string(char *data, int maxlen, int write,
2001			   char __user *buffer,
2002			   size_t *lenp, loff_t *ppos)
2003{
2004	size_t len;
2005	char __user *p;
2006	char c;
2007
2008	if (!data || !maxlen || !*lenp) {
2009		*lenp = 0;
2010		return 0;
2011	}
2012
2013	if (write) {
2014		if (sysctl_writes_strict == SYSCTL_WRITES_STRICT) {
2015			/* Only continue writes not past the end of buffer. */
2016			len = strlen(data);
2017			if (len > maxlen - 1)
2018				len = maxlen - 1;
2019
2020			if (*ppos > len)
2021				return 0;
2022			len = *ppos;
2023		} else {
2024			/* Start writing from beginning of buffer. */
2025			len = 0;
2026		}
2027
2028		*ppos += *lenp;
2029		p = buffer;
2030		while ((p - buffer) < *lenp && len < maxlen - 1) {
2031			if (get_user(c, p++))
2032				return -EFAULT;
2033			if (c == 0 || c == '\n')
2034				break;
2035			data[len++] = c;
2036		}
2037		data[len] = 0;
2038	} else {
2039		len = strlen(data);
2040		if (len > maxlen)
2041			len = maxlen;
2042
2043		if (*ppos > len) {
2044			*lenp = 0;
2045			return 0;
2046		}
2047
2048		data += *ppos;
2049		len  -= *ppos;
2050
2051		if (len > *lenp)
2052			len = *lenp;
2053		if (len)
2054			if (copy_to_user(buffer, data, len))
2055				return -EFAULT;
2056		if (len < *lenp) {
2057			if (put_user('\n', buffer + len))
2058				return -EFAULT;
2059			len++;
2060		}
2061		*lenp = len;
2062		*ppos += len;
2063	}
2064	return 0;
2065}
2066
2067static void warn_sysctl_write(struct ctl_table *table)
2068{
2069	pr_warn_once("%s wrote to %s when file position was not 0!\n"
2070		"This will not be supported in the future. To silence this\n"
2071		"warning, set kernel.sysctl_writes_strict = -1\n",
2072		current->comm, table->procname);
2073}
2074
2075/**
2076 * proc_first_pos_non_zero_ignore - check if first position is allowed
2077 * @ppos: file position
2078 * @table: the sysctl table
2079 *
2080 * Returns true if the first position is non-zero and the sysctl_writes_strict
2081 * mode indicates this is not allowed for numeric input types. String proc
2082 * handlers can ignore the return value.
2083 */
2084static bool proc_first_pos_non_zero_ignore(loff_t *ppos,
2085					   struct ctl_table *table)
2086{
2087	if (!*ppos)
2088		return false;
2089
2090	switch (sysctl_writes_strict) {
2091	case SYSCTL_WRITES_STRICT:
2092		return true;
2093	case SYSCTL_WRITES_WARN:
2094		warn_sysctl_write(table);
2095		return false;
2096	default:
2097		return false;
2098	}
2099}
2100
2101/**
2102 * proc_dostring - read a string sysctl
2103 * @table: the sysctl table
2104 * @write: %TRUE if this is a write to the sysctl file
2105 * @buffer: the user buffer
2106 * @lenp: the size of the user buffer
2107 * @ppos: file position
2108 *
2109 * Reads/writes a string from/to the user buffer. If the kernel
2110 * buffer provided is not large enough to hold the string, the
2111 * string is truncated. The copied string is %NULL-terminated.
2112 * If the string is being read by the user process, it is copied
2113 * and a newline '\n' is added. It is truncated if the buffer is
2114 * not large enough.
2115 *
2116 * Returns 0 on success.
2117 */
2118int proc_dostring(struct ctl_table *table, int write,
2119		  void __user *buffer, size_t *lenp, loff_t *ppos)
2120{
2121	if (write)
2122		proc_first_pos_non_zero_ignore(ppos, table);
2123
2124	return _proc_do_string((char *)(table->data), table->maxlen, write,
2125			       (char __user *)buffer, lenp, ppos);
2126}
2127
2128static size_t proc_skip_spaces(char **buf)
2129{
2130	size_t ret;
2131	char *tmp = skip_spaces(*buf);
2132	ret = tmp - *buf;
2133	*buf = tmp;
2134	return ret;
2135}
2136
2137static void proc_skip_char(char **buf, size_t *size, const char v)
2138{
2139	while (*size) {
2140		if (**buf != v)
2141			break;
2142		(*size)--;
2143		(*buf)++;
2144	}
2145}
2146
2147/**
2148 * strtoul_lenient - parse an ASCII formatted integer from a buffer and only
2149 *                   fail on overflow
2150 *
2151 * @cp: kernel buffer containing the string to parse
2152 * @endp: pointer to store the trailing characters
2153 * @base: the base to use
2154 * @res: where the parsed integer will be stored
2155 *
2156 * In case of success 0 is returned and @res will contain the parsed integer,
2157 * @endp will hold any trailing characters.
2158 * This function will fail the parse on overflow. If there wasn't an overflow
2159 * the function will defer the decision what characters count as invalid to the
2160 * caller.
2161 */
2162static int strtoul_lenient(const char *cp, char **endp, unsigned int base,
2163			   unsigned long *res)
2164{
2165	unsigned long long result;
2166	unsigned int rv;
2167
2168	cp = _parse_integer_fixup_radix(cp, &base);
2169	rv = _parse_integer(cp, base, &result);
2170	if ((rv & KSTRTOX_OVERFLOW) || (result != (unsigned long)result))
2171		return -ERANGE;
2172
2173	cp += rv;
2174
2175	if (endp)
2176		*endp = (char *)cp;
2177
2178	*res = (unsigned long)result;
2179	return 0;
2180}
2181
2182#define TMPBUFLEN 22
2183/**
2184 * proc_get_long - reads an ASCII formatted integer from a user buffer
2185 *
2186 * @buf: a kernel buffer
2187 * @size: size of the kernel buffer
2188 * @val: this is where the number will be stored
2189 * @neg: set to %TRUE if number is negative
2190 * @perm_tr: a vector which contains the allowed trailers
2191 * @perm_tr_len: size of the perm_tr vector
2192 * @tr: pointer to store the trailer character
2193 *
2194 * In case of success %0 is returned and @buf and @size are updated with
2195 * the amount of bytes read. If @tr is non-NULL and a trailing
2196 * character exists (size is non-zero after returning from this
2197 * function), @tr is updated with the trailing character.
2198 */
2199static int proc_get_long(char **buf, size_t *size,
2200			  unsigned long *val, bool *neg,
2201			  const char *perm_tr, unsigned perm_tr_len, char *tr)
2202{
2203	int len;
2204	char *p, tmp[TMPBUFLEN];
2205
2206	if (!*size)
2207		return -EINVAL;
2208
2209	len = *size;
2210	if (len > TMPBUFLEN - 1)
2211		len = TMPBUFLEN - 1;
2212
2213	memcpy(tmp, *buf, len);
2214
2215	tmp[len] = 0;
2216	p = tmp;
2217	if (*p == '-' && *size > 1) {
2218		*neg = true;
2219		p++;
2220	} else
2221		*neg = false;
2222	if (!isdigit(*p))
2223		return -EINVAL;
2224
2225	if (strtoul_lenient(p, &p, 0, val))
2226		return -EINVAL;
2227
2228	len = p - tmp;
2229
2230	/* We don't know if the next char is whitespace thus we may accept
2231	 * invalid integers (e.g. 1234...a) or two integers instead of one
2232	 * (e.g. 123...1). So lets not allow such large numbers. */
2233	if (len == TMPBUFLEN - 1)
2234		return -EINVAL;
2235
2236	if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
2237		return -EINVAL;
2238
2239	if (tr && (len < *size))
2240		*tr = *p;
2241
2242	*buf += len;
2243	*size -= len;
2244
2245	return 0;
2246}
2247
2248/**
2249 * proc_put_long - converts an integer to a decimal ASCII formatted string
2250 *
2251 * @buf: the user buffer
2252 * @size: the size of the user buffer
2253 * @val: the integer to be converted
2254 * @neg: sign of the number, %TRUE for negative
2255 *
2256 * In case of success %0 is returned and @buf and @size are updated with
2257 * the amount of bytes written.
2258 */
2259static int proc_put_long(void __user **buf, size_t *size, unsigned long val,
2260			  bool neg)
2261{
2262	int len;
2263	char tmp[TMPBUFLEN], *p = tmp;
2264
2265	sprintf(p, "%s%lu", neg ? "-" : "", val);
2266	len = strlen(tmp);
2267	if (len > *size)
2268		len = *size;
2269	if (copy_to_user(*buf, tmp, len))
2270		return -EFAULT;
2271	*size -= len;
2272	*buf += len;
2273	return 0;
2274}
2275#undef TMPBUFLEN
2276
2277static int proc_put_char(void __user **buf, size_t *size, char c)
2278{
2279	if (*size) {
2280		char __user **buffer = (char __user **)buf;
2281		if (put_user(c, *buffer))
2282			return -EFAULT;
2283		(*size)--, (*buffer)++;
2284		*buf = *buffer;
2285	}
2286	return 0;
2287}
2288
2289static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
2290				 int *valp,
2291				 int write, void *data)
2292{
2293	if (write) {
2294		if (*negp) {
2295			if (*lvalp > (unsigned long) INT_MAX + 1)
2296				return -EINVAL;
2297			*valp = -*lvalp;
2298		} else {
2299			if (*lvalp > (unsigned long) INT_MAX)
2300				return -EINVAL;
2301			*valp = *lvalp;
2302		}
2303	} else {
2304		int val = *valp;
2305		if (val < 0) {
2306			*negp = true;
2307			*lvalp = -(unsigned long)val;
2308		} else {
2309			*negp = false;
2310			*lvalp = (unsigned long)val;
2311		}
2312	}
2313	return 0;
2314}
2315
2316static int do_proc_douintvec_conv(unsigned long *lvalp,
2317				  unsigned int *valp,
2318				  int write, void *data)
2319{
2320	if (write) {
2321		if (*lvalp > UINT_MAX)
2322			return -EINVAL;
2323		*valp = *lvalp;
2324	} else {
2325		unsigned int val = *valp;
2326		*lvalp = (unsigned long)val;
2327	}
2328	return 0;
2329}
2330
2331static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
2332
2333static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
2334		  int write, void __user *buffer,
2335		  size_t *lenp, loff_t *ppos,
2336		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
2337			      int write, void *data),
2338		  void *data)
2339{
2340	int *i, vleft, first = 1, err = 0;
2341	size_t left;
2342	char *kbuf = NULL, *p;
2343	
2344	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
2345		*lenp = 0;
2346		return 0;
2347	}
2348	
2349	i = (int *) tbl_data;
2350	vleft = table->maxlen / sizeof(*i);
2351	left = *lenp;
2352
2353	if (!conv)
2354		conv = do_proc_dointvec_conv;
2355
2356	if (write) {
2357		if (proc_first_pos_non_zero_ignore(ppos, table))
2358			goto out;
2359
2360		if (left > PAGE_SIZE - 1)
2361			left = PAGE_SIZE - 1;
2362		p = kbuf = memdup_user_nul(buffer, left);
2363		if (IS_ERR(kbuf))
2364			return PTR_ERR(kbuf);
2365	}
2366
2367	for (; left && vleft--; i++, first=0) {
2368		unsigned long lval;
2369		bool neg;
2370
2371		if (write) {
2372			left -= proc_skip_spaces(&p);
2373
2374			if (!left)
2375				break;
2376			err = proc_get_long(&p, &left, &lval, &neg,
2377					     proc_wspace_sep,
2378					     sizeof(proc_wspace_sep), NULL);
2379			if (err)
2380				break;
2381			if (conv(&neg, &lval, i, 1, data)) {
2382				err = -EINVAL;
2383				break;
2384			}
2385		} else {
2386			if (conv(&neg, &lval, i, 0, data)) {
2387				err = -EINVAL;
2388				break;
2389			}
2390			if (!first)
2391				err = proc_put_char(&buffer, &left, '\t');
2392			if (err)
2393				break;
2394			err = proc_put_long(&buffer, &left, lval, neg);
2395			if (err)
2396				break;
2397		}
2398	}
2399
2400	if (!write && !first && left && !err)
2401		err = proc_put_char(&buffer, &left, '\n');
2402	if (write && !err && left)
2403		left -= proc_skip_spaces(&p);
2404	if (write) {
2405		kfree(kbuf);
2406		if (first)
2407			return err ? : -EINVAL;
2408	}
2409	*lenp -= left;
2410out:
2411	*ppos += *lenp;
2412	return err;
2413}
2414
2415static int do_proc_dointvec(struct ctl_table *table, int write,
2416		  void __user *buffer, size_t *lenp, loff_t *ppos,
2417		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
2418			      int write, void *data),
2419		  void *data)
2420{
2421	return __do_proc_dointvec(table->data, table, write,
2422			buffer, lenp, ppos, conv, data);
2423}
2424
2425static int do_proc_douintvec_w(unsigned int *tbl_data,
2426			       struct ctl_table *table,
2427			       void __user *buffer,
2428			       size_t *lenp, loff_t *ppos,
2429			       int (*conv)(unsigned long *lvalp,
2430					   unsigned int *valp,
2431					   int write, void *data),
2432			       void *data)
2433{
2434	unsigned long lval;
2435	int err = 0;
2436	size_t left;
2437	bool neg;
2438	char *kbuf = NULL, *p;
2439
2440	left = *lenp;
2441
2442	if (proc_first_pos_non_zero_ignore(ppos, table))
2443		goto bail_early;
2444
2445	if (left > PAGE_SIZE - 1)
2446		left = PAGE_SIZE - 1;
2447
2448	p = kbuf = memdup_user_nul(buffer, left);
2449	if (IS_ERR(kbuf))
2450		return -EINVAL;
2451
2452	left -= proc_skip_spaces(&p);
2453	if (!left) {
2454		err = -EINVAL;
2455		goto out_free;
2456	}
2457
2458	err = proc_get_long(&p, &left, &lval, &neg,
2459			     proc_wspace_sep,
2460			     sizeof(proc_wspace_sep), NULL);
2461	if (err || neg) {
2462		err = -EINVAL;
2463		goto out_free;
2464	}
2465
2466	if (conv(&lval, tbl_data, 1, data)) {
2467		err = -EINVAL;
2468		goto out_free;
2469	}
2470
2471	if (!err && left)
2472		left -= proc_skip_spaces(&p);
2473
2474out_free:
2475	kfree(kbuf);
2476	if (err)
2477		return -EINVAL;
2478
2479	return 0;
2480
2481	/* This is in keeping with old __do_proc_dointvec() */
2482bail_early:
2483	*ppos += *lenp;
2484	return err;
2485}
2486
2487static int do_proc_douintvec_r(unsigned int *tbl_data, void __user *buffer,
2488			       size_t *lenp, loff_t *ppos,
2489			       int (*conv)(unsigned long *lvalp,
2490					   unsigned int *valp,
2491					   int write, void *data),
2492			       void *data)
2493{
2494	unsigned long lval;
2495	int err = 0;
2496	size_t left;
2497
2498	left = *lenp;
2499
2500	if (conv(&lval, tbl_data, 0, data)) {
2501		err = -EINVAL;
2502		goto out;
2503	}
2504
2505	err = proc_put_long(&buffer, &left, lval, false);
2506	if (err || !left)
2507		goto out;
2508
2509	err = proc_put_char(&buffer, &left, '\n');
2510
2511out:
2512	*lenp -= left;
2513	*ppos += *lenp;
2514
2515	return err;
2516}
2517
2518static int __do_proc_douintvec(void *tbl_data, struct ctl_table *table,
2519			       int write, void __user *buffer,
2520			       size_t *lenp, loff_t *ppos,
2521			       int (*conv)(unsigned long *lvalp,
2522					   unsigned int *valp,
2523					   int write, void *data),
2524			       void *data)
2525{
2526	unsigned int *i, vleft;
2527
2528	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
2529		*lenp = 0;
2530		return 0;
2531	}
2532
2533	i = (unsigned int *) tbl_data;
2534	vleft = table->maxlen / sizeof(*i);
2535
2536	/*
2537	 * Arrays are not supported, keep this simple. *Do not* add
2538	 * support for them.
2539	 */
2540	if (vleft != 1) {
2541		*lenp = 0;
2542		return -EINVAL;
2543	}
2544
2545	if (!conv)
2546		conv = do_proc_douintvec_conv;
2547
2548	if (write)
2549		return do_proc_douintvec_w(i, table, buffer, lenp, ppos,
2550					   conv, data);
2551	return do_proc_douintvec_r(i, buffer, lenp, ppos, conv, data);
2552}
2553
2554static int do_proc_douintvec(struct ctl_table *table, int write,
2555			     void __user *buffer, size_t *lenp, loff_t *ppos,
2556			     int (*conv)(unsigned long *lvalp,
2557					 unsigned int *valp,
2558					 int write, void *data),
2559			     void *data)
2560{
2561	return __do_proc_douintvec(table->data, table, write,
2562				   buffer, lenp, ppos, conv, data);
2563}
2564
2565/**
2566 * proc_dointvec - read a vector of integers
2567 * @table: the sysctl table
2568 * @write: %TRUE if this is a write to the sysctl file
2569 * @buffer: the user buffer
2570 * @lenp: the size of the user buffer
2571 * @ppos: file position
2572 *
2573 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2574 * values from/to the user buffer, treated as an ASCII string. 
2575 *
2576 * Returns 0 on success.
2577 */
2578int proc_dointvec(struct ctl_table *table, int write,
2579		     void __user *buffer, size_t *lenp, loff_t *ppos)
2580{
2581	return do_proc_dointvec(table, write, buffer, lenp, ppos, NULL, NULL);
2582}
2583
2584/**
2585 * proc_douintvec - read a vector of unsigned integers
2586 * @table: the sysctl table
2587 * @write: %TRUE if this is a write to the sysctl file
2588 * @buffer: the user buffer
2589 * @lenp: the size of the user buffer
2590 * @ppos: file position
2591 *
2592 * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
2593 * values from/to the user buffer, treated as an ASCII string.
2594 *
2595 * Returns 0 on success.
2596 */
2597int proc_douintvec(struct ctl_table *table, int write,
2598		     void __user *buffer, size_t *lenp, loff_t *ppos)
2599{
2600	return do_proc_douintvec(table, write, buffer, lenp, ppos,
2601				 do_proc_douintvec_conv, NULL);
2602}
2603
2604/*
2605 * Taint values can only be increased
2606 * This means we can safely use a temporary.
2607 */
2608static int proc_taint(struct ctl_table *table, int write,
2609			       void __user *buffer, size_t *lenp, loff_t *ppos)
2610{
2611	struct ctl_table t;
2612	unsigned long tmptaint = get_taint();
2613	int err;
2614
2615	if (write && !capable(CAP_SYS_ADMIN))
2616		return -EPERM;
2617
2618	t = *table;
2619	t.data = &tmptaint;
2620	err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
2621	if (err < 0)
2622		return err;
2623
2624	if (write) {
2625		/*
2626		 * Poor man's atomic or. Not worth adding a primitive
2627		 * to everyone's atomic.h for this
2628		 */
2629		int i;
2630		for (i = 0; i < BITS_PER_LONG && tmptaint >> i; i++) {
2631			if ((tmptaint >> i) & 1)
2632				add_taint(i, LOCKDEP_STILL_OK);
2633		}
2634	}
2635
2636	return err;
2637}
2638
2639#ifdef CONFIG_PRINTK
2640static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
2641				void __user *buffer, size_t *lenp, loff_t *ppos)
2642{
2643	if (write && !capable(CAP_SYS_ADMIN))
2644		return -EPERM;
2645
2646	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2647}
2648#endif
2649
2650/**
2651 * struct do_proc_dointvec_minmax_conv_param - proc_dointvec_minmax() range checking structure
2652 * @min: pointer to minimum allowable value
2653 * @max: pointer to maximum allowable value
2654 *
2655 * The do_proc_dointvec_minmax_conv_param structure provides the
2656 * minimum and maximum values for doing range checking for those sysctl
2657 * parameters that use the proc_dointvec_minmax() handler.
2658 */
2659struct do_proc_dointvec_minmax_conv_param {
2660	int *min;
2661	int *max;
2662};
2663
2664static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
2665					int *valp,
2666					int write, void *data)
2667{
2668	int tmp, ret;
2669	struct do_proc_dointvec_minmax_conv_param *param = data;
2670	/*
2671	 * If writing, first do so via a temporary local int so we can
2672	 * bounds-check it before touching *valp.
2673	 */
2674	int *ip = write ? &tmp : valp;
2675
2676	ret = do_proc_dointvec_conv(negp, lvalp, ip, write, data);
2677	if (ret)
2678		return ret;
2679
2680	if (write) {
2681		if ((param->min && *param->min > tmp) ||
2682		    (param->max && *param->max < tmp))
2683			return -EINVAL;
2684		*valp = tmp;
2685	}
2686
2687	return 0;
2688}
2689
2690/**
2691 * proc_dointvec_minmax - read a vector of integers with min/max values
2692 * @table: the sysctl table
2693 * @write: %TRUE if this is a write to the sysctl file
2694 * @buffer: the user buffer
2695 * @lenp: the size of the user buffer
2696 * @ppos: file position
2697 *
2698 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2699 * values from/to the user buffer, treated as an ASCII string.
2700 *
2701 * This routine will ensure the values are within the range specified by
2702 * table->extra1 (min) and table->extra2 (max).
2703 *
2704 * Returns 0 on success or -EINVAL on write when the range check fails.
2705 */
2706int proc_dointvec_minmax(struct ctl_table *table, int write,
2707		  void __user *buffer, size_t *lenp, loff_t *ppos)
2708{
2709	struct do_proc_dointvec_minmax_conv_param param = {
2710		.min = (int *) table->extra1,
2711		.max = (int *) table->extra2,
2712	};
2713	return do_proc_dointvec(table, write, buffer, lenp, ppos,
2714				do_proc_dointvec_minmax_conv, &param);
2715}
2716
2717/**
2718 * struct do_proc_douintvec_minmax_conv_param - proc_douintvec_minmax() range checking structure
2719 * @min: pointer to minimum allowable value
2720 * @max: pointer to maximum allowable value
2721 *
2722 * The do_proc_douintvec_minmax_conv_param structure provides the
2723 * minimum and maximum values for doing range checking for those sysctl
2724 * parameters that use the proc_douintvec_minmax() handler.
2725 */
2726struct do_proc_douintvec_minmax_conv_param {
2727	unsigned int *min;
2728	unsigned int *max;
2729};
2730
2731static int do_proc_douintvec_minmax_conv(unsigned long *lvalp,
2732					 unsigned int *valp,
2733					 int write, void *data)
2734{
2735	int ret;
2736	unsigned int tmp;
2737	struct do_proc_douintvec_minmax_conv_param *param = data;
2738	/* write via temporary local uint for bounds-checking */
2739	unsigned int *up = write ? &tmp : valp;
2740
2741	ret = do_proc_douintvec_conv(lvalp, up, write, data);
2742	if (ret)
2743		return ret;
2744
2745	if (write) {
2746		if ((param->min && *param->min > tmp) ||
2747		    (param->max && *param->max < tmp))
2748			return -ERANGE;
2749
2750		*valp = tmp;
2751	}
2752
2753	return 0;
2754}
2755
2756/**
2757 * proc_douintvec_minmax - read a vector of unsigned ints with min/max values
2758 * @table: the sysctl table
2759 * @write: %TRUE if this is a write to the sysctl file
2760 * @buffer: the user buffer
2761 * @lenp: the size of the user buffer
2762 * @ppos: file position
2763 *
2764 * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
2765 * values from/to the user buffer, treated as an ASCII string. Negative
2766 * strings are not allowed.
2767 *
2768 * This routine will ensure the values are within the range specified by
2769 * table->extra1 (min) and table->extra2 (max). There is a final sanity
2770 * check for UINT_MAX to avoid having to support wrap around uses from
2771 * userspace.
2772 *
2773 * Returns 0 on success or -ERANGE on write when the range check fails.
2774 */
2775int proc_douintvec_minmax(struct ctl_table *table, int write,
2776			  void __user *buffer, size_t *lenp, loff_t *ppos)
2777{
2778	struct do_proc_douintvec_minmax_conv_param param = {
2779		.min = (unsigned int *) table->extra1,
2780		.max = (unsigned int *) table->extra2,
2781	};
2782	return do_proc_douintvec(table, write, buffer, lenp, ppos,
2783				 do_proc_douintvec_minmax_conv, &param);
2784}
2785
2786static int do_proc_dopipe_max_size_conv(unsigned long *lvalp,
2787					unsigned int *valp,
2788					int write, void *data)
2789{
2790	if (write) {
2791		unsigned int val;
2792
2793		val = round_pipe_size(*lvalp);
2794		if (val == 0)
2795			return -EINVAL;
2796
2797		*valp = val;
2798	} else {
2799		unsigned int val = *valp;
2800		*lvalp = (unsigned long) val;
2801	}
2802
2803	return 0;
2804}
2805
2806static int proc_dopipe_max_size(struct ctl_table *table, int write,
2807				void __user *buffer, size_t *lenp, loff_t *ppos)
2808{
2809	return do_proc_douintvec(table, write, buffer, lenp, ppos,
2810				 do_proc_dopipe_max_size_conv, NULL);
2811}
2812
2813static void validate_coredump_safety(void)
2814{
2815#ifdef CONFIG_COREDUMP
2816	if (suid_dumpable == SUID_DUMP_ROOT &&
2817	    core_pattern[0] != '/' && core_pattern[0] != '|') {
2818		printk(KERN_WARNING
2819"Unsafe core_pattern used with fs.suid_dumpable=2.\n"
2820"Pipe handler or fully qualified core dump path required.\n"
2821"Set kernel.core_pattern before fs.suid_dumpable.\n"
2822		);
2823	}
2824#endif
2825}
2826
2827static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
2828		void __user *buffer, size_t *lenp, loff_t *ppos)
2829{
2830	int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2831	if (!error)
2832		validate_coredump_safety();
2833	return error;
2834}
2835
2836#ifdef CONFIG_COREDUMP
2837static int proc_dostring_coredump(struct ctl_table *table, int write,
2838		  void __user *buffer, size_t *lenp, loff_t *ppos)
2839{
2840	int error = proc_dostring(table, write, buffer, lenp, ppos);
2841	if (!error)
2842		validate_coredump_safety();
2843	return error;
2844}
2845#endif
2846
2847static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table, int write,
2848				     void __user *buffer,
2849				     size_t *lenp, loff_t *ppos,
2850				     unsigned long convmul,
2851				     unsigned long convdiv)
2852{
2853	unsigned long *i, *min, *max;
2854	int vleft, first = 1, err = 0;
2855	size_t left;
2856	char *kbuf = NULL, *p;
2857
2858	if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
2859		*lenp = 0;
2860		return 0;
2861	}
2862
2863	i = (unsigned long *) data;
2864	min = (unsigned long *) table->extra1;
2865	max = (unsigned long *) table->extra2;
2866	vleft = table->maxlen / sizeof(unsigned long);
2867	left = *lenp;
2868
2869	if (write) {
2870		if (proc_first_pos_non_zero_ignore(ppos, table))
2871			goto out;
2872
2873		if (left > PAGE_SIZE - 1)
2874			left = PAGE_SIZE - 1;
2875		p = kbuf = memdup_user_nul(buffer, left);
2876		if (IS_ERR(kbuf))
2877			return PTR_ERR(kbuf);
2878	}
2879
2880	for (; left && vleft--; i++, first = 0) {
2881		unsigned long val;
2882
2883		if (write) {
2884			bool neg;
2885
2886			left -= proc_skip_spaces(&p);
2887			if (!left)
2888				break;
2889
2890			err = proc_get_long(&p, &left, &val, &neg,
2891					     proc_wspace_sep,
2892					     sizeof(proc_wspace_sep), NULL);
2893			if (err)
2894				break;
2895			if (neg)
2896				continue;
2897			val = convmul * val / convdiv;
2898			if ((min && val < *min) || (max && val > *max)) {
2899				err = -EINVAL;
2900				break;
2901			}
2902			*i = val;
2903		} else {
2904			val = convdiv * (*i) / convmul;
2905			if (!first) {
2906				err = proc_put_char(&buffer, &left, '\t');
2907				if (err)
2908					break;
2909			}
2910			err = proc_put_long(&buffer, &left, val, false);
2911			if (err)
2912				break;
2913		}
2914	}
2915
2916	if (!write && !first && left && !err)
2917		err = proc_put_char(&buffer, &left, '\n');
2918	if (write && !err)
2919		left -= proc_skip_spaces(&p);
2920	if (write) {
2921		kfree(kbuf);
2922		if (first)
2923			return err ? : -EINVAL;
2924	}
2925	*lenp -= left;
2926out:
2927	*ppos += *lenp;
2928	return err;
2929}
2930
2931static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
2932				     void __user *buffer,
2933				     size_t *lenp, loff_t *ppos,
2934				     unsigned long convmul,
2935				     unsigned long convdiv)
2936{
2937	return __do_proc_doulongvec_minmax(table->data, table, write,
2938			buffer, lenp, ppos, convmul, convdiv);
2939}
2940
2941/**
2942 * proc_doulongvec_minmax - read a vector of long integers with min/max values
2943 * @table: the sysctl table
2944 * @write: %TRUE if this is a write to the sysctl file
2945 * @buffer: the user buffer
2946 * @lenp: the size of the user buffer
2947 * @ppos: file position
2948 *
2949 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2950 * values from/to the user buffer, treated as an ASCII string.
2951 *
2952 * This routine will ensure the values are within the range specified by
2953 * table->extra1 (min) and table->extra2 (max).
2954 *
2955 * Returns 0 on success.
2956 */
2957int proc_doulongvec_minmax(struct ctl_table *table, int write,
2958			   void __user *buffer, size_t *lenp, loff_t *ppos)
2959{
2960    return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
2961}
2962
2963/**
2964 * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
2965 * @table: the sysctl table
2966 * @write: %TRUE if this is a write to the sysctl file
2967 * @buffer: the user buffer
2968 * @lenp: the size of the user buffer
2969 * @ppos: file position
2970 *
2971 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2972 * values from/to the user buffer, treated as an ASCII string. The values
2973 * are treated as milliseconds, and converted to jiffies when they are stored.
2974 *
2975 * This routine will ensure the values are within the range specified by
2976 * table->extra1 (min) and table->extra2 (max).
2977 *
2978 * Returns 0 on success.
2979 */
2980int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
2981				      void __user *buffer,
2982				      size_t *lenp, loff_t *ppos)
2983{
2984    return do_proc_doulongvec_minmax(table, write, buffer,
2985				     lenp, ppos, HZ, 1000l);
2986}
2987
2988
2989static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
2990					 int *valp,
2991					 int write, void *data)
2992{
2993	if (write) {
2994		if (*lvalp > INT_MAX / HZ)
2995			return 1;
2996		*valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
2997	} else {
2998		int val = *valp;
2999		unsigned long lval;
3000		if (val < 0) {
3001			*negp = true;
3002			lval = -(unsigned long)val;
3003		} else {
3004			*negp = false;
3005			lval = (unsigned long)val;
3006		}
3007		*lvalp = lval / HZ;
3008	}
3009	return 0;
3010}
3011
3012static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
3013						int *valp,
3014						int write, void *data)
3015{
3016	if (write) {
3017		if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
3018			return 1;
3019		*valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
3020	} else {
3021		int val = *valp;
3022		unsigned long lval;
3023		if (val < 0) {
3024			*negp = true;
3025			lval = -(unsigned long)val;
3026		} else {
3027			*negp = false;
3028			lval = (unsigned long)val;
3029		}
3030		*lvalp = jiffies_to_clock_t(lval);
3031	}
3032	return 0;
3033}
3034
3035static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
3036					    int *valp,
3037					    int write, void *data)
3038{
3039	if (write) {
3040		unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
3041
3042		if (jif > INT_MAX)
3043			return 1;
3044		*valp = (int)jif;
3045	} else {
3046		int val = *valp;
3047		unsigned long lval;
3048		if (val < 0) {
3049			*negp = true;
3050			lval = -(unsigned long)val;
3051		} else {
3052			*negp = false;
3053			lval = (unsigned long)val;
3054		}
3055		*lvalp = jiffies_to_msecs(lval);
3056	}
3057	return 0;
3058}
3059
3060/**
3061 * proc_dointvec_jiffies - read a vector of integers as seconds
3062 * @table: the sysctl table
3063 * @write: %TRUE if this is a write to the sysctl file
3064 * @buffer: the user buffer
3065 * @lenp: the size of the user buffer
3066 * @ppos: file position
3067 *
3068 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
3069 * values from/to the user buffer, treated as an ASCII string. 
3070 * The values read are assumed to be in seconds, and are converted into
3071 * jiffies.
3072 *
3073 * Returns 0 on success.
3074 */
3075int proc_dointvec_jiffies(struct ctl_table *table, int write,
3076			  void __user *buffer, size_t *lenp, loff_t *ppos)
3077{
3078    return do_proc_dointvec(table,write,buffer,lenp,ppos,
3079		    	    do_proc_dointvec_jiffies_conv,NULL);
3080}
3081
3082/**
3083 * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
3084 * @table: the sysctl table
3085 * @write: %TRUE if this is a write to the sysctl file
3086 * @buffer: the user buffer
3087 * @lenp: the size of the user buffer
3088 * @ppos: pointer to the file position
3089 *
3090 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
3091 * values from/to the user buffer, treated as an ASCII string. 
3092 * The values read are assumed to be in 1/USER_HZ seconds, and 
3093 * are converted into jiffies.
3094 *
3095 * Returns 0 on success.
3096 */
3097int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
3098				 void __user *buffer, size_t *lenp, loff_t *ppos)
3099{
3100    return do_proc_dointvec(table,write,buffer,lenp,ppos,
3101		    	    do_proc_dointvec_userhz_jiffies_conv,NULL);
3102}
3103
3104/**
3105 * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
3106 * @table: the sysctl table
3107 * @write: %TRUE if this is a write to the sysctl file
3108 * @buffer: the user buffer
3109 * @lenp: the size of the user buffer
3110 * @ppos: file position
3111 * @ppos: the current position in the file
3112 *
3113 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
3114 * values from/to the user buffer, treated as an ASCII string. 
3115 * The values read are assumed to be in 1/1000 seconds, and 
3116 * are converted into jiffies.
3117 *
3118 * Returns 0 on success.
3119 */
3120int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
3121			     void __user *buffer, size_t *lenp, loff_t *ppos)
3122{
3123	return do_proc_dointvec(table, write, buffer, lenp, ppos,
3124				do_proc_dointvec_ms_jiffies_conv, NULL);
3125}
3126
3127static int proc_do_cad_pid(struct ctl_table *table, int write,
3128			   void __user *buffer, size_t *lenp, loff_t *ppos)
3129{
3130	struct pid *new_pid;
3131	pid_t tmp;
3132	int r;
3133
3134	tmp = pid_vnr(cad_pid);
3135
3136	r = __do_proc_dointvec(&tmp, table, write, buffer,
3137			       lenp, ppos, NULL, NULL);
3138	if (r || !write)
3139		return r;
3140
3141	new_pid = find_get_pid(tmp);
3142	if (!new_pid)
3143		return -ESRCH;
3144
3145	put_pid(xchg(&cad_pid, new_pid));
3146	return 0;
3147}
3148
3149/**
3150 * proc_do_large_bitmap - read/write from/to a large bitmap
3151 * @table: the sysctl table
3152 * @write: %TRUE if this is a write to the sysctl file
3153 * @buffer: the user buffer
3154 * @lenp: the size of the user buffer
3155 * @ppos: file position
3156 *
3157 * The bitmap is stored at table->data and the bitmap length (in bits)
3158 * in table->maxlen.
3159 *
3160 * We use a range comma separated format (e.g. 1,3-4,10-10) so that
3161 * large bitmaps may be represented in a compact manner. Writing into
3162 * the file will clear the bitmap then update it with the given input.
3163 *
3164 * Returns 0 on success.
3165 */
3166int proc_do_large_bitmap(struct ctl_table *table, int write,
3167			 void __user *buffer, size_t *lenp, loff_t *ppos)
3168{
3169	int err = 0;
3170	bool first = 1;
3171	size_t left = *lenp;
3172	unsigned long bitmap_len = table->maxlen;
3173	unsigned long *bitmap = *(unsigned long **) table->data;
3174	unsigned long *tmp_bitmap = NULL;
3175	char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
3176
3177	if (!bitmap || !bitmap_len || !left || (*ppos && !write)) {
3178		*lenp = 0;
3179		return 0;
3180	}
3181
3182	if (write) {
3183		char *kbuf, *p;
3184		size_t skipped = 0;
3185
3186		if (left > PAGE_SIZE - 1) {
3187			left = PAGE_SIZE - 1;
3188			/* How much of the buffer we'll skip this pass */
3189			skipped = *lenp - left;
3190		}
3191
3192		p = kbuf = memdup_user_nul(buffer, left);
3193		if (IS_ERR(kbuf))
3194			return PTR_ERR(kbuf);
3195
3196		tmp_bitmap = bitmap_zalloc(bitmap_len, GFP_KERNEL);
3197		if (!tmp_bitmap) {
3198			kfree(kbuf);
3199			return -ENOMEM;
3200		}
3201		proc_skip_char(&p, &left, '\n');
3202		while (!err && left) {
3203			unsigned long val_a, val_b;
3204			bool neg;
3205			size_t saved_left;
3206
3207			/* In case we stop parsing mid-number, we can reset */
3208			saved_left = left;
3209			err = proc_get_long(&p, &left, &val_a, &neg, tr_a,
3210					     sizeof(tr_a), &c);
3211			/*
3212			 * If we consumed the entirety of a truncated buffer or
3213			 * only one char is left (may be a "-"), then stop here,
3214			 * reset, & come back for more.
3215			 */
3216			if ((left <= 1) && skipped) {
3217				left = saved_left;
3218				break;
3219			}
3220
3221			if (err)
3222				break;
3223			if (val_a >= bitmap_len || neg) {
3224				err = -EINVAL;
3225				break;
3226			}
3227
3228			val_b = val_a;
3229			if (left) {
3230				p++;
3231				left--;
3232			}
3233
3234			if (c == '-') {
3235				err = proc_get_long(&p, &left, &val_b,
3236						     &neg, tr_b, sizeof(tr_b),
3237						     &c);
3238				/*
3239				 * If we consumed all of a truncated buffer or
3240				 * then stop here, reset, & come back for more.
3241				 */
3242				if (!left && skipped) {
3243					left = saved_left;
3244					break;
3245				}
3246
3247				if (err)
3248					break;
3249				if (val_b >= bitmap_len || neg ||
3250				    val_a > val_b) {
3251					err = -EINVAL;
3252					break;
3253				}
3254				if (left) {
3255					p++;
3256					left--;
3257				}
3258			}
3259
3260			bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
3261			first = 0;
3262			proc_skip_char(&p, &left, '\n');
3263		}
3264		kfree(kbuf);
3265		left += skipped;
3266	} else {
3267		unsigned long bit_a, bit_b = 0;
3268
3269		while (left) {
3270			bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
3271			if (bit_a >= bitmap_len)
3272				break;
3273			bit_b = find_next_zero_bit(bitmap, bitmap_len,
3274						   bit_a + 1) - 1;
3275
3276			if (!first) {
3277				err = proc_put_char(&buffer, &left, ',');
3278				if (err)
3279					break;
3280			}
3281			err = proc_put_long(&buffer, &left, bit_a, false);
3282			if (err)
3283				break;
3284			if (bit_a != bit_b) {
3285				err = proc_put_char(&buffer, &left, '-');
3286				if (err)
3287					break;
3288				err = proc_put_long(&buffer, &left, bit_b, false);
3289				if (err)
3290					break;
3291			}
3292
3293			first = 0; bit_b++;
3294		}
3295		if (!err)
3296			err = proc_put_char(&buffer, &left, '\n');
3297	}
3298
3299	if (!err) {
3300		if (write) {
3301			if (*ppos)
3302				bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
3303			else
3304				bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
3305		}
3306		*lenp -= left;
3307		*ppos += *lenp;
3308	}
3309
3310	bitmap_free(tmp_bitmap);
3311	return err;
3312}
3313
3314#else /* CONFIG_PROC_SYSCTL */
3315
3316int proc_dostring(struct ctl_table *table, int write,
3317		  void __user *buffer, size_t *lenp, loff_t *ppos)
3318{
3319	return -ENOSYS;
3320}
3321
3322int proc_dointvec(struct ctl_table *table, int write,
3323		  void __user *buffer, size_t *lenp, loff_t *ppos)
3324{
3325	return -ENOSYS;
3326}
3327
3328int proc_douintvec(struct ctl_table *table, int write,
3329		  void __user *buffer, size_t *lenp, loff_t *ppos)
3330{
3331	return -ENOSYS;
3332}
3333
3334int proc_dointvec_minmax(struct ctl_table *table, int write,
3335		    void __user *buffer, size_t *lenp, loff_t *ppos)
3336{
3337	return -ENOSYS;
3338}
3339
3340int proc_douintvec_minmax(struct ctl_table *table, int write,
3341			  void __user *buffer, size_t *lenp, loff_t *ppos)
3342{
3343	return -ENOSYS;
3344}
3345
3346int proc_dointvec_jiffies(struct ctl_table *table, int write,
3347		    void __user *buffer, size_t *lenp, loff_t *ppos)
3348{
3349	return -ENOSYS;
3350}
3351
3352int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
3353		    void __user *buffer, size_t *lenp, loff_t *ppos)
3354{
3355	return -ENOSYS;
3356}
3357
3358int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
3359			     void __user *buffer, size_t *lenp, loff_t *ppos)
3360{
3361	return -ENOSYS;
3362}
3363
3364int proc_doulongvec_minmax(struct ctl_table *table, int write,
3365		    void __user *buffer, size_t *lenp, loff_t *ppos)
3366{
3367	return -ENOSYS;
3368}
3369
3370int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
3371				      void __user *buffer,
3372				      size_t *lenp, loff_t *ppos)
3373{
3374    return -ENOSYS;
3375}
3376
3377int proc_do_large_bitmap(struct ctl_table *table, int write,
3378			 void __user *buffer, size_t *lenp, loff_t *ppos)
3379{
3380	return -ENOSYS;
3381}
3382
3383#endif /* CONFIG_PROC_SYSCTL */
3384
3385#if defined(CONFIG_SYSCTL)
3386int proc_do_static_key(struct ctl_table *table, int write,
3387		       void __user *buffer, size_t *lenp,
3388		       loff_t *ppos)
3389{
3390	struct static_key *key = (struct static_key *)table->data;
3391	static DEFINE_MUTEX(static_key_mutex);
3392	int val, ret;
3393	struct ctl_table tmp = {
3394		.data   = &val,
3395		.maxlen = sizeof(val),
3396		.mode   = table->mode,
3397		.extra1 = SYSCTL_ZERO,
3398		.extra2 = SYSCTL_ONE,
3399	};
3400
3401	if (write && !capable(CAP_SYS_ADMIN))
3402		return -EPERM;
3403
3404	mutex_lock(&static_key_mutex);
3405	val = static_key_enabled(key);
3406	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
3407	if (write && !ret) {
3408		if (val)
3409			static_key_enable(key);
3410		else
3411			static_key_disable(key);
3412	}
3413	mutex_unlock(&static_key_mutex);
3414	return ret;
3415}
3416#endif
3417/*
3418 * No sense putting this after each symbol definition, twice,
3419 * exception granted :-)
3420 */
3421EXPORT_SYMBOL(proc_dointvec);
3422EXPORT_SYMBOL(proc_douintvec);
3423EXPORT_SYMBOL(proc_dointvec_jiffies);
3424EXPORT_SYMBOL(proc_dointvec_minmax);
3425EXPORT_SYMBOL_GPL(proc_douintvec_minmax);
3426EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
3427EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
3428EXPORT_SYMBOL(proc_dostring);
3429EXPORT_SYMBOL(proc_doulongvec_minmax);
3430EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
3431EXPORT_SYMBOL(proc_do_large_bitmap);
v5.9
   1// SPDX-License-Identifier: GPL-2.0-only
   2/*
   3 * sysctl.c: General linux system control interface
   4 *
   5 * Begun 24 March 1995, Stephen Tweedie
   6 * Added /proc support, Dec 1995
   7 * Added bdflush entry and intvec min/max checking, 2/23/96, Tom Dyas.
   8 * Added hooks for /proc/sys/net (minor, minor patch), 96/4/1, Mike Shaver.
   9 * Added kernel/java-{interpreter,appletviewer}, 96/5/10, Mike Shaver.
  10 * Dynamic registration fixes, Stephen Tweedie.
  11 * Added kswapd-interval, ctrl-alt-del, printk stuff, 1/8/97, Chris Horn.
  12 * Made sysctl support optional via CONFIG_SYSCTL, 1/10/97, Chris
  13 *  Horn.
  14 * Added proc_doulongvec_ms_jiffies_minmax, 09/08/99, Carlos H. Bauer.
  15 * Added proc_doulongvec_minmax, 09/08/99, Carlos H. Bauer.
  16 * Changed linked lists to use list.h instead of lists.h, 02/24/00, Bill
  17 *  Wendling.
  18 * The list_for_each() macro wasn't appropriate for the sysctl loop.
  19 *  Removed it and replaced it with older style, 03/23/00, Bill Wendling
  20 */
  21
  22#include <linux/module.h>
  23#include <linux/aio.h>
  24#include <linux/mm.h>
  25#include <linux/swap.h>
  26#include <linux/slab.h>
  27#include <linux/sysctl.h>
  28#include <linux/bitmap.h>
  29#include <linux/signal.h>
  30#include <linux/printk.h>
  31#include <linux/proc_fs.h>
  32#include <linux/security.h>
  33#include <linux/ctype.h>
  34#include <linux/kmemleak.h>
  35#include <linux/fs.h>
  36#include <linux/init.h>
  37#include <linux/kernel.h>
  38#include <linux/kobject.h>
  39#include <linux/net.h>
  40#include <linux/sysrq.h>
  41#include <linux/highuid.h>
  42#include <linux/writeback.h>
  43#include <linux/ratelimit.h>
  44#include <linux/compaction.h>
  45#include <linux/hugetlb.h>
  46#include <linux/initrd.h>
  47#include <linux/key.h>
  48#include <linux/times.h>
  49#include <linux/limits.h>
  50#include <linux/dcache.h>
  51#include <linux/dnotify.h>
  52#include <linux/syscalls.h>
  53#include <linux/vmstat.h>
  54#include <linux/nfs_fs.h>
  55#include <linux/acpi.h>
  56#include <linux/reboot.h>
  57#include <linux/ftrace.h>
  58#include <linux/perf_event.h>
  59#include <linux/kprobes.h>
  60#include <linux/pipe_fs_i.h>
  61#include <linux/oom.h>
  62#include <linux/kmod.h>
  63#include <linux/capability.h>
  64#include <linux/binfmts.h>
  65#include <linux/sched/sysctl.h>
  66#include <linux/sched/coredump.h>
  67#include <linux/kexec.h>
  68#include <linux/bpf.h>
  69#include <linux/mount.h>
  70#include <linux/userfaultfd_k.h>
  71#include <linux/coredump.h>
  72#include <linux/latencytop.h>
  73#include <linux/pid.h>
  74
  75#include "../lib/kstrtox.h"
  76
  77#include <linux/uaccess.h>
  78#include <asm/processor.h>
  79
  80#ifdef CONFIG_X86
  81#include <asm/nmi.h>
  82#include <asm/stacktrace.h>
  83#include <asm/io.h>
  84#endif
  85#ifdef CONFIG_SPARC
  86#include <asm/setup.h>
  87#endif
  88#ifdef CONFIG_BSD_PROCESS_ACCT
  89#include <linux/acct.h>
  90#endif
  91#ifdef CONFIG_RT_MUTEXES
  92#include <linux/rtmutex.h>
  93#endif
  94#if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
  95#include <linux/lockdep.h>
  96#endif
  97#ifdef CONFIG_CHR_DEV_SG
  98#include <scsi/sg.h>
  99#endif
 100#ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
 101#include <linux/stackleak.h>
 102#endif
 103#ifdef CONFIG_LOCKUP_DETECTOR
 104#include <linux/nmi.h>
 105#endif
 106
 107#if defined(CONFIG_SYSCTL)
 108
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 109/* Constants used for minimum and  maximum */
 110#ifdef CONFIG_LOCKUP_DETECTOR
 111static int sixty = 60;
 112#endif
 113
 114static int __maybe_unused neg_one = -1;
 115static int __maybe_unused two = 2;
 116static int __maybe_unused four = 4;
 117static unsigned long zero_ul;
 118static unsigned long one_ul = 1;
 119static unsigned long long_max = LONG_MAX;
 120static int one_hundred = 100;
 121static int two_hundred = 200;
 122static int one_thousand = 1000;
 123#ifdef CONFIG_PRINTK
 124static int ten_thousand = 10000;
 125#endif
 126#ifdef CONFIG_PERF_EVENTS
 127static int six_hundred_forty_kb = 640 * 1024;
 128#endif
 129
 130/* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
 131static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
 132
 133/* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
 134static int maxolduid = 65535;
 135static int minolduid;
 136
 137static int ngroups_max = NGROUPS_MAX;
 138static const int cap_last_cap = CAP_LAST_CAP;
 139
 140/*
 141 * This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
 142 * and hung_task_check_interval_secs
 143 */
 144#ifdef CONFIG_DETECT_HUNG_TASK
 145static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
 146#endif
 147
 148#ifdef CONFIG_INOTIFY_USER
 149#include <linux/inotify.h>
 150#endif
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 151
 152#ifdef CONFIG_PROC_SYSCTL
 153
 154/**
 155 * enum sysctl_writes_mode - supported sysctl write modes
 156 *
 157 * @SYSCTL_WRITES_LEGACY: each write syscall must fully contain the sysctl value
 158 *	to be written, and multiple writes on the same sysctl file descriptor
 159 *	will rewrite the sysctl value, regardless of file position. No warning
 160 *	is issued when the initial position is not 0.
 161 * @SYSCTL_WRITES_WARN: same as above but warn when the initial file position is
 162 *	not 0.
 163 * @SYSCTL_WRITES_STRICT: writes to numeric sysctl entries must always be at
 164 *	file position 0 and the value must be fully contained in the buffer
 165 *	sent to the write syscall. If dealing with strings respect the file
 166 *	position, but restrict this to the max length of the buffer, anything
 167 *	passed the max length will be ignored. Multiple writes will append
 168 *	to the buffer.
 169 *
 170 * These write modes control how current file position affects the behavior of
 171 * updating sysctl values through the proc interface on each write.
 172 */
 173enum sysctl_writes_mode {
 174	SYSCTL_WRITES_LEGACY		= -1,
 175	SYSCTL_WRITES_WARN		= 0,
 176	SYSCTL_WRITES_STRICT		= 1,
 177};
 178
 179static enum sysctl_writes_mode sysctl_writes_strict = SYSCTL_WRITES_STRICT;
 180#endif /* CONFIG_PROC_SYSCTL */
 181
 182#if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
 183    defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
 184int sysctl_legacy_va_layout;
 185#endif
 186
 187#ifdef CONFIG_SCHED_DEBUG
 188static int min_sched_granularity_ns = 100000;		/* 100 usecs */
 189static int max_sched_granularity_ns = NSEC_PER_SEC;	/* 1 second */
 190static int min_wakeup_granularity_ns;			/* 0 usecs */
 191static int max_wakeup_granularity_ns = NSEC_PER_SEC;	/* 1 second */
 192#ifdef CONFIG_SMP
 193static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
 194static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
 195#endif /* CONFIG_SMP */
 196#endif /* CONFIG_SCHED_DEBUG */
 197
 198#ifdef CONFIG_COMPACTION
 199static int min_extfrag_threshold;
 200static int max_extfrag_threshold = 1000;
 201#endif
 202
 203#endif /* CONFIG_SYSCTL */
 204
 205#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_SYSCTL)
 206static int bpf_stats_handler(struct ctl_table *table, int write,
 207			     void *buffer, size_t *lenp, loff_t *ppos)
 208{
 209	struct static_key *key = (struct static_key *)table->data;
 210	static int saved_val;
 211	int val, ret;
 212	struct ctl_table tmp = {
 213		.data   = &val,
 214		.maxlen = sizeof(val),
 215		.mode   = table->mode,
 216		.extra1 = SYSCTL_ZERO,
 217		.extra2 = SYSCTL_ONE,
 218	};
 219
 220	if (write && !capable(CAP_SYS_ADMIN))
 221		return -EPERM;
 222
 223	mutex_lock(&bpf_stats_enabled_mutex);
 224	val = saved_val;
 225	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
 226	if (write && !ret && val != saved_val) {
 227		if (val)
 228			static_key_slow_inc(key);
 229		else
 230			static_key_slow_dec(key);
 231		saved_val = val;
 232	}
 233	mutex_unlock(&bpf_stats_enabled_mutex);
 234	return ret;
 235}
 236#endif
 237
 238/*
 239 * /proc/sys support
 240 */
 241
 242#ifdef CONFIG_PROC_SYSCTL
 243
 244static int _proc_do_string(char *data, int maxlen, int write,
 245		char *buffer, size_t *lenp, loff_t *ppos)
 246{
 247	size_t len;
 248	char c, *p;
 249
 250	if (!data || !maxlen || !*lenp) {
 251		*lenp = 0;
 252		return 0;
 253	}
 254
 255	if (write) {
 256		if (sysctl_writes_strict == SYSCTL_WRITES_STRICT) {
 257			/* Only continue writes not past the end of buffer. */
 258			len = strlen(data);
 259			if (len > maxlen - 1)
 260				len = maxlen - 1;
 261
 262			if (*ppos > len)
 263				return 0;
 264			len = *ppos;
 265		} else {
 266			/* Start writing from beginning of buffer. */
 267			len = 0;
 268		}
 269
 270		*ppos += *lenp;
 271		p = buffer;
 272		while ((p - buffer) < *lenp && len < maxlen - 1) {
 273			c = *(p++);
 274			if (c == 0 || c == '\n')
 275				break;
 276			data[len++] = c;
 277		}
 278		data[len] = 0;
 279	} else {
 280		len = strlen(data);
 281		if (len > maxlen)
 282			len = maxlen;
 283
 284		if (*ppos > len) {
 285			*lenp = 0;
 286			return 0;
 287		}
 288
 289		data += *ppos;
 290		len  -= *ppos;
 291
 292		if (len > *lenp)
 293			len = *lenp;
 294		if (len)
 295			memcpy(buffer, data, len);
 296		if (len < *lenp) {
 297			buffer[len] = '\n';
 298			len++;
 299		}
 300		*lenp = len;
 301		*ppos += len;
 302	}
 303	return 0;
 304}
 305
 306static void warn_sysctl_write(struct ctl_table *table)
 307{
 308	pr_warn_once("%s wrote to %s when file position was not 0!\n"
 309		"This will not be supported in the future. To silence this\n"
 310		"warning, set kernel.sysctl_writes_strict = -1\n",
 311		current->comm, table->procname);
 312}
 313
 314/**
 315 * proc_first_pos_non_zero_ignore - check if first position is allowed
 316 * @ppos: file position
 317 * @table: the sysctl table
 318 *
 319 * Returns true if the first position is non-zero and the sysctl_writes_strict
 320 * mode indicates this is not allowed for numeric input types. String proc
 321 * handlers can ignore the return value.
 322 */
 323static bool proc_first_pos_non_zero_ignore(loff_t *ppos,
 324					   struct ctl_table *table)
 325{
 326	if (!*ppos)
 327		return false;
 328
 329	switch (sysctl_writes_strict) {
 330	case SYSCTL_WRITES_STRICT:
 331		return true;
 332	case SYSCTL_WRITES_WARN:
 333		warn_sysctl_write(table);
 334		return false;
 335	default:
 336		return false;
 337	}
 338}
 339
 340/**
 341 * proc_dostring - read a string sysctl
 342 * @table: the sysctl table
 343 * @write: %TRUE if this is a write to the sysctl file
 344 * @buffer: the user buffer
 345 * @lenp: the size of the user buffer
 346 * @ppos: file position
 347 *
 348 * Reads/writes a string from/to the user buffer. If the kernel
 349 * buffer provided is not large enough to hold the string, the
 350 * string is truncated. The copied string is %NULL-terminated.
 351 * If the string is being read by the user process, it is copied
 352 * and a newline '\n' is added. It is truncated if the buffer is
 353 * not large enough.
 354 *
 355 * Returns 0 on success.
 356 */
 357int proc_dostring(struct ctl_table *table, int write,
 358		  void *buffer, size_t *lenp, loff_t *ppos)
 359{
 360	if (write)
 361		proc_first_pos_non_zero_ignore(ppos, table);
 362
 363	return _proc_do_string(table->data, table->maxlen, write, buffer, lenp,
 364			ppos);
 365}
 366
 367static size_t proc_skip_spaces(char **buf)
 368{
 369	size_t ret;
 370	char *tmp = skip_spaces(*buf);
 371	ret = tmp - *buf;
 372	*buf = tmp;
 373	return ret;
 374}
 375
 376static void proc_skip_char(char **buf, size_t *size, const char v)
 377{
 378	while (*size) {
 379		if (**buf != v)
 380			break;
 381		(*size)--;
 382		(*buf)++;
 383	}
 384}
 385
 386/**
 387 * strtoul_lenient - parse an ASCII formatted integer from a buffer and only
 388 *                   fail on overflow
 389 *
 390 * @cp: kernel buffer containing the string to parse
 391 * @endp: pointer to store the trailing characters
 392 * @base: the base to use
 393 * @res: where the parsed integer will be stored
 394 *
 395 * In case of success 0 is returned and @res will contain the parsed integer,
 396 * @endp will hold any trailing characters.
 397 * This function will fail the parse on overflow. If there wasn't an overflow
 398 * the function will defer the decision what characters count as invalid to the
 399 * caller.
 400 */
 401static int strtoul_lenient(const char *cp, char **endp, unsigned int base,
 402			   unsigned long *res)
 403{
 404	unsigned long long result;
 405	unsigned int rv;
 406
 407	cp = _parse_integer_fixup_radix(cp, &base);
 408	rv = _parse_integer(cp, base, &result);
 409	if ((rv & KSTRTOX_OVERFLOW) || (result != (unsigned long)result))
 410		return -ERANGE;
 411
 412	cp += rv;
 413
 414	if (endp)
 415		*endp = (char *)cp;
 416
 417	*res = (unsigned long)result;
 418	return 0;
 419}
 420
 421#define TMPBUFLEN 22
 422/**
 423 * proc_get_long - reads an ASCII formatted integer from a user buffer
 424 *
 425 * @buf: a kernel buffer
 426 * @size: size of the kernel buffer
 427 * @val: this is where the number will be stored
 428 * @neg: set to %TRUE if number is negative
 429 * @perm_tr: a vector which contains the allowed trailers
 430 * @perm_tr_len: size of the perm_tr vector
 431 * @tr: pointer to store the trailer character
 432 *
 433 * In case of success %0 is returned and @buf and @size are updated with
 434 * the amount of bytes read. If @tr is non-NULL and a trailing
 435 * character exists (size is non-zero after returning from this
 436 * function), @tr is updated with the trailing character.
 437 */
 438static int proc_get_long(char **buf, size_t *size,
 439			  unsigned long *val, bool *neg,
 440			  const char *perm_tr, unsigned perm_tr_len, char *tr)
 441{
 442	int len;
 443	char *p, tmp[TMPBUFLEN];
 444
 445	if (!*size)
 446		return -EINVAL;
 447
 448	len = *size;
 449	if (len > TMPBUFLEN - 1)
 450		len = TMPBUFLEN - 1;
 451
 452	memcpy(tmp, *buf, len);
 453
 454	tmp[len] = 0;
 455	p = tmp;
 456	if (*p == '-' && *size > 1) {
 457		*neg = true;
 458		p++;
 459	} else
 460		*neg = false;
 461	if (!isdigit(*p))
 462		return -EINVAL;
 463
 464	if (strtoul_lenient(p, &p, 0, val))
 465		return -EINVAL;
 466
 467	len = p - tmp;
 468
 469	/* We don't know if the next char is whitespace thus we may accept
 470	 * invalid integers (e.g. 1234...a) or two integers instead of one
 471	 * (e.g. 123...1). So lets not allow such large numbers. */
 472	if (len == TMPBUFLEN - 1)
 473		return -EINVAL;
 474
 475	if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
 476		return -EINVAL;
 477
 478	if (tr && (len < *size))
 479		*tr = *p;
 480
 481	*buf += len;
 482	*size -= len;
 483
 484	return 0;
 485}
 486
 487/**
 488 * proc_put_long - converts an integer to a decimal ASCII formatted string
 489 *
 490 * @buf: the user buffer
 491 * @size: the size of the user buffer
 492 * @val: the integer to be converted
 493 * @neg: sign of the number, %TRUE for negative
 494 *
 495 * In case of success @buf and @size are updated with the amount of bytes
 496 * written.
 497 */
 498static void proc_put_long(void **buf, size_t *size, unsigned long val, bool neg)
 499{
 500	int len;
 501	char tmp[TMPBUFLEN], *p = tmp;
 502
 503	sprintf(p, "%s%lu", neg ? "-" : "", val);
 504	len = strlen(tmp);
 505	if (len > *size)
 506		len = *size;
 507	memcpy(*buf, tmp, len);
 508	*size -= len;
 509	*buf += len;
 510}
 511#undef TMPBUFLEN
 512
 513static void proc_put_char(void **buf, size_t *size, char c)
 514{
 515	if (*size) {
 516		char **buffer = (char **)buf;
 517		**buffer = c;
 518
 519		(*size)--;
 520		(*buffer)++;
 521		*buf = *buffer;
 522	}
 523}
 524
 525static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
 526				 int *valp,
 527				 int write, void *data)
 528{
 529	if (write) {
 530		if (*negp) {
 531			if (*lvalp > (unsigned long) INT_MAX + 1)
 532				return -EINVAL;
 533			*valp = -*lvalp;
 534		} else {
 535			if (*lvalp > (unsigned long) INT_MAX)
 536				return -EINVAL;
 537			*valp = *lvalp;
 538		}
 539	} else {
 540		int val = *valp;
 541		if (val < 0) {
 542			*negp = true;
 543			*lvalp = -(unsigned long)val;
 544		} else {
 545			*negp = false;
 546			*lvalp = (unsigned long)val;
 547		}
 548	}
 549	return 0;
 550}
 551
 552static int do_proc_douintvec_conv(unsigned long *lvalp,
 553				  unsigned int *valp,
 554				  int write, void *data)
 555{
 556	if (write) {
 557		if (*lvalp > UINT_MAX)
 558			return -EINVAL;
 559		*valp = *lvalp;
 560	} else {
 561		unsigned int val = *valp;
 562		*lvalp = (unsigned long)val;
 563	}
 564	return 0;
 565}
 566
 567static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
 568
 569static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
 570		  int write, void *buffer,
 571		  size_t *lenp, loff_t *ppos,
 572		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
 573			      int write, void *data),
 574		  void *data)
 575{
 576	int *i, vleft, first = 1, err = 0;
 577	size_t left;
 578	char *p;
 579	
 580	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
 581		*lenp = 0;
 582		return 0;
 583	}
 584	
 585	i = (int *) tbl_data;
 586	vleft = table->maxlen / sizeof(*i);
 587	left = *lenp;
 588
 589	if (!conv)
 590		conv = do_proc_dointvec_conv;
 591
 592	if (write) {
 593		if (proc_first_pos_non_zero_ignore(ppos, table))
 594			goto out;
 595
 596		if (left > PAGE_SIZE - 1)
 597			left = PAGE_SIZE - 1;
 598		p = buffer;
 599	}
 600
 601	for (; left && vleft--; i++, first=0) {
 602		unsigned long lval;
 603		bool neg;
 604
 605		if (write) {
 606			left -= proc_skip_spaces(&p);
 607
 608			if (!left)
 609				break;
 610			err = proc_get_long(&p, &left, &lval, &neg,
 611					     proc_wspace_sep,
 612					     sizeof(proc_wspace_sep), NULL);
 613			if (err)
 614				break;
 615			if (conv(&neg, &lval, i, 1, data)) {
 616				err = -EINVAL;
 617				break;
 618			}
 619		} else {
 620			if (conv(&neg, &lval, i, 0, data)) {
 621				err = -EINVAL;
 622				break;
 623			}
 624			if (!first)
 625				proc_put_char(&buffer, &left, '\t');
 626			proc_put_long(&buffer, &left, lval, neg);
 627		}
 628	}
 629
 630	if (!write && !first && left && !err)
 631		proc_put_char(&buffer, &left, '\n');
 632	if (write && !err && left)
 633		left -= proc_skip_spaces(&p);
 634	if (write && first)
 635		return err ? : -EINVAL;
 636	*lenp -= left;
 637out:
 638	*ppos += *lenp;
 639	return err;
 640}
 641
 642static int do_proc_dointvec(struct ctl_table *table, int write,
 643		  void *buffer, size_t *lenp, loff_t *ppos,
 644		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
 645			      int write, void *data),
 646		  void *data)
 647{
 648	return __do_proc_dointvec(table->data, table, write,
 649			buffer, lenp, ppos, conv, data);
 650}
 651
 652static int do_proc_douintvec_w(unsigned int *tbl_data,
 653			       struct ctl_table *table,
 654			       void *buffer,
 655			       size_t *lenp, loff_t *ppos,
 656			       int (*conv)(unsigned long *lvalp,
 657					   unsigned int *valp,
 658					   int write, void *data),
 659			       void *data)
 660{
 661	unsigned long lval;
 662	int err = 0;
 663	size_t left;
 664	bool neg;
 665	char *p = buffer;
 666
 667	left = *lenp;
 668
 669	if (proc_first_pos_non_zero_ignore(ppos, table))
 670		goto bail_early;
 671
 672	if (left > PAGE_SIZE - 1)
 673		left = PAGE_SIZE - 1;
 674
 675	left -= proc_skip_spaces(&p);
 676	if (!left) {
 677		err = -EINVAL;
 678		goto out_free;
 679	}
 680
 681	err = proc_get_long(&p, &left, &lval, &neg,
 682			     proc_wspace_sep,
 683			     sizeof(proc_wspace_sep), NULL);
 684	if (err || neg) {
 685		err = -EINVAL;
 686		goto out_free;
 687	}
 688
 689	if (conv(&lval, tbl_data, 1, data)) {
 690		err = -EINVAL;
 691		goto out_free;
 692	}
 693
 694	if (!err && left)
 695		left -= proc_skip_spaces(&p);
 696
 697out_free:
 698	if (err)
 699		return -EINVAL;
 700
 701	return 0;
 702
 703	/* This is in keeping with old __do_proc_dointvec() */
 704bail_early:
 705	*ppos += *lenp;
 706	return err;
 707}
 708
 709static int do_proc_douintvec_r(unsigned int *tbl_data, void *buffer,
 710			       size_t *lenp, loff_t *ppos,
 711			       int (*conv)(unsigned long *lvalp,
 712					   unsigned int *valp,
 713					   int write, void *data),
 714			       void *data)
 715{
 716	unsigned long lval;
 717	int err = 0;
 718	size_t left;
 719
 720	left = *lenp;
 721
 722	if (conv(&lval, tbl_data, 0, data)) {
 723		err = -EINVAL;
 724		goto out;
 725	}
 726
 727	proc_put_long(&buffer, &left, lval, false);
 728	if (!left)
 729		goto out;
 730
 731	proc_put_char(&buffer, &left, '\n');
 732
 733out:
 734	*lenp -= left;
 735	*ppos += *lenp;
 736
 737	return err;
 738}
 739
 740static int __do_proc_douintvec(void *tbl_data, struct ctl_table *table,
 741			       int write, void *buffer,
 742			       size_t *lenp, loff_t *ppos,
 743			       int (*conv)(unsigned long *lvalp,
 744					   unsigned int *valp,
 745					   int write, void *data),
 746			       void *data)
 747{
 748	unsigned int *i, vleft;
 749
 750	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
 751		*lenp = 0;
 752		return 0;
 753	}
 754
 755	i = (unsigned int *) tbl_data;
 756	vleft = table->maxlen / sizeof(*i);
 757
 758	/*
 759	 * Arrays are not supported, keep this simple. *Do not* add
 760	 * support for them.
 761	 */
 762	if (vleft != 1) {
 763		*lenp = 0;
 764		return -EINVAL;
 765	}
 766
 767	if (!conv)
 768		conv = do_proc_douintvec_conv;
 769
 770	if (write)
 771		return do_proc_douintvec_w(i, table, buffer, lenp, ppos,
 772					   conv, data);
 773	return do_proc_douintvec_r(i, buffer, lenp, ppos, conv, data);
 774}
 775
 776static int do_proc_douintvec(struct ctl_table *table, int write,
 777			     void *buffer, size_t *lenp, loff_t *ppos,
 778			     int (*conv)(unsigned long *lvalp,
 779					 unsigned int *valp,
 780					 int write, void *data),
 781			     void *data)
 782{
 783	return __do_proc_douintvec(table->data, table, write,
 784				   buffer, lenp, ppos, conv, data);
 785}
 786
 787/**
 788 * proc_dointvec - read a vector of integers
 789 * @table: the sysctl table
 790 * @write: %TRUE if this is a write to the sysctl file
 791 * @buffer: the user buffer
 792 * @lenp: the size of the user buffer
 793 * @ppos: file position
 794 *
 795 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
 796 * values from/to the user buffer, treated as an ASCII string. 
 797 *
 798 * Returns 0 on success.
 799 */
 800int proc_dointvec(struct ctl_table *table, int write, void *buffer,
 801		  size_t *lenp, loff_t *ppos)
 802{
 803	return do_proc_dointvec(table, write, buffer, lenp, ppos, NULL, NULL);
 804}
 805
 806#ifdef CONFIG_COMPACTION
 807static int proc_dointvec_minmax_warn_RT_change(struct ctl_table *table,
 808		int write, void *buffer, size_t *lenp, loff_t *ppos)
 809{
 810	int ret, old;
 811
 812	if (!IS_ENABLED(CONFIG_PREEMPT_RT) || !write)
 813		return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
 814
 815	old = *(int *)table->data;
 816	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
 817	if (ret)
 818		return ret;
 819	if (old != *(int *)table->data)
 820		pr_warn_once("sysctl attribute %s changed by %s[%d]\n",
 821			     table->procname, current->comm,
 822			     task_pid_nr(current));
 823	return ret;
 824}
 825#endif
 826
 827/**
 828 * proc_douintvec - read a vector of unsigned integers
 829 * @table: the sysctl table
 830 * @write: %TRUE if this is a write to the sysctl file
 831 * @buffer: the user buffer
 832 * @lenp: the size of the user buffer
 833 * @ppos: file position
 834 *
 835 * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
 836 * values from/to the user buffer, treated as an ASCII string.
 837 *
 838 * Returns 0 on success.
 839 */
 840int proc_douintvec(struct ctl_table *table, int write, void *buffer,
 841		size_t *lenp, loff_t *ppos)
 842{
 843	return do_proc_douintvec(table, write, buffer, lenp, ppos,
 844				 do_proc_douintvec_conv, NULL);
 845}
 846
 847/*
 848 * Taint values can only be increased
 849 * This means we can safely use a temporary.
 850 */
 851static int proc_taint(struct ctl_table *table, int write,
 852			       void *buffer, size_t *lenp, loff_t *ppos)
 853{
 854	struct ctl_table t;
 855	unsigned long tmptaint = get_taint();
 856	int err;
 857
 858	if (write && !capable(CAP_SYS_ADMIN))
 859		return -EPERM;
 860
 861	t = *table;
 862	t.data = &tmptaint;
 863	err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
 864	if (err < 0)
 865		return err;
 866
 867	if (write) {
 868		int i;
 869
 870		/*
 871		 * If we are relying on panic_on_taint not producing
 872		 * false positives due to userspace input, bail out
 873		 * before setting the requested taint flags.
 874		 */
 875		if (panic_on_taint_nousertaint && (tmptaint & panic_on_taint))
 876			return -EINVAL;
 877
 878		/*
 879		 * Poor man's atomic or. Not worth adding a primitive
 880		 * to everyone's atomic.h for this
 881		 */
 882		for (i = 0; i < TAINT_FLAGS_COUNT; i++)
 883			if ((1UL << i) & tmptaint)
 884				add_taint(i, LOCKDEP_STILL_OK);
 885	}
 886
 887	return err;
 888}
 889
 890#ifdef CONFIG_PRINTK
 891static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
 892				void *buffer, size_t *lenp, loff_t *ppos)
 893{
 894	if (write && !capable(CAP_SYS_ADMIN))
 895		return -EPERM;
 896
 897	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
 898}
 899#endif
 900
 901/**
 902 * struct do_proc_dointvec_minmax_conv_param - proc_dointvec_minmax() range checking structure
 903 * @min: pointer to minimum allowable value
 904 * @max: pointer to maximum allowable value
 905 *
 906 * The do_proc_dointvec_minmax_conv_param structure provides the
 907 * minimum and maximum values for doing range checking for those sysctl
 908 * parameters that use the proc_dointvec_minmax() handler.
 909 */
 910struct do_proc_dointvec_minmax_conv_param {
 911	int *min;
 912	int *max;
 913};
 914
 915static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
 916					int *valp,
 917					int write, void *data)
 918{
 919	int tmp, ret;
 920	struct do_proc_dointvec_minmax_conv_param *param = data;
 921	/*
 922	 * If writing, first do so via a temporary local int so we can
 923	 * bounds-check it before touching *valp.
 924	 */
 925	int *ip = write ? &tmp : valp;
 926
 927	ret = do_proc_dointvec_conv(negp, lvalp, ip, write, data);
 928	if (ret)
 929		return ret;
 930
 931	if (write) {
 932		if ((param->min && *param->min > tmp) ||
 933		    (param->max && *param->max < tmp))
 934			return -EINVAL;
 935		*valp = tmp;
 936	}
 937
 938	return 0;
 939}
 940
 941/**
 942 * proc_dointvec_minmax - read a vector of integers with min/max values
 943 * @table: the sysctl table
 944 * @write: %TRUE if this is a write to the sysctl file
 945 * @buffer: the user buffer
 946 * @lenp: the size of the user buffer
 947 * @ppos: file position
 948 *
 949 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
 950 * values from/to the user buffer, treated as an ASCII string.
 951 *
 952 * This routine will ensure the values are within the range specified by
 953 * table->extra1 (min) and table->extra2 (max).
 954 *
 955 * Returns 0 on success or -EINVAL on write when the range check fails.
 956 */
 957int proc_dointvec_minmax(struct ctl_table *table, int write,
 958		  void *buffer, size_t *lenp, loff_t *ppos)
 959{
 960	struct do_proc_dointvec_minmax_conv_param param = {
 961		.min = (int *) table->extra1,
 962		.max = (int *) table->extra2,
 963	};
 964	return do_proc_dointvec(table, write, buffer, lenp, ppos,
 965				do_proc_dointvec_minmax_conv, &param);
 966}
 967
 968/**
 969 * struct do_proc_douintvec_minmax_conv_param - proc_douintvec_minmax() range checking structure
 970 * @min: pointer to minimum allowable value
 971 * @max: pointer to maximum allowable value
 972 *
 973 * The do_proc_douintvec_minmax_conv_param structure provides the
 974 * minimum and maximum values for doing range checking for those sysctl
 975 * parameters that use the proc_douintvec_minmax() handler.
 976 */
 977struct do_proc_douintvec_minmax_conv_param {
 978	unsigned int *min;
 979	unsigned int *max;
 980};
 981
 982static int do_proc_douintvec_minmax_conv(unsigned long *lvalp,
 983					 unsigned int *valp,
 984					 int write, void *data)
 985{
 986	int ret;
 987	unsigned int tmp;
 988	struct do_proc_douintvec_minmax_conv_param *param = data;
 989	/* write via temporary local uint for bounds-checking */
 990	unsigned int *up = write ? &tmp : valp;
 991
 992	ret = do_proc_douintvec_conv(lvalp, up, write, data);
 993	if (ret)
 994		return ret;
 995
 996	if (write) {
 997		if ((param->min && *param->min > tmp) ||
 998		    (param->max && *param->max < tmp))
 999			return -ERANGE;
1000
1001		*valp = tmp;
1002	}
1003
1004	return 0;
1005}
1006
1007/**
1008 * proc_douintvec_minmax - read a vector of unsigned ints with min/max values
1009 * @table: the sysctl table
1010 * @write: %TRUE if this is a write to the sysctl file
1011 * @buffer: the user buffer
1012 * @lenp: the size of the user buffer
1013 * @ppos: file position
1014 *
1015 * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
1016 * values from/to the user buffer, treated as an ASCII string. Negative
1017 * strings are not allowed.
1018 *
1019 * This routine will ensure the values are within the range specified by
1020 * table->extra1 (min) and table->extra2 (max). There is a final sanity
1021 * check for UINT_MAX to avoid having to support wrap around uses from
1022 * userspace.
1023 *
1024 * Returns 0 on success or -ERANGE on write when the range check fails.
1025 */
1026int proc_douintvec_minmax(struct ctl_table *table, int write,
1027			  void *buffer, size_t *lenp, loff_t *ppos)
1028{
1029	struct do_proc_douintvec_minmax_conv_param param = {
1030		.min = (unsigned int *) table->extra1,
1031		.max = (unsigned int *) table->extra2,
1032	};
1033	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1034				 do_proc_douintvec_minmax_conv, &param);
1035}
1036
1037static int do_proc_dopipe_max_size_conv(unsigned long *lvalp,
1038					unsigned int *valp,
1039					int write, void *data)
1040{
1041	if (write) {
1042		unsigned int val;
1043
1044		val = round_pipe_size(*lvalp);
1045		if (val == 0)
1046			return -EINVAL;
1047
1048		*valp = val;
1049	} else {
1050		unsigned int val = *valp;
1051		*lvalp = (unsigned long) val;
1052	}
1053
1054	return 0;
1055}
1056
1057static int proc_dopipe_max_size(struct ctl_table *table, int write,
1058				void *buffer, size_t *lenp, loff_t *ppos)
1059{
1060	return do_proc_douintvec(table, write, buffer, lenp, ppos,
1061				 do_proc_dopipe_max_size_conv, NULL);
1062}
1063
1064static void validate_coredump_safety(void)
1065{
1066#ifdef CONFIG_COREDUMP
1067	if (suid_dumpable == SUID_DUMP_ROOT &&
1068	    core_pattern[0] != '/' && core_pattern[0] != '|') {
1069		printk(KERN_WARNING
1070"Unsafe core_pattern used with fs.suid_dumpable=2.\n"
1071"Pipe handler or fully qualified core dump path required.\n"
1072"Set kernel.core_pattern before fs.suid_dumpable.\n"
1073		);
1074	}
1075#endif
1076}
1077
1078static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
1079		void *buffer, size_t *lenp, loff_t *ppos)
1080{
1081	int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
1082	if (!error)
1083		validate_coredump_safety();
1084	return error;
1085}
1086
1087#ifdef CONFIG_COREDUMP
1088static int proc_dostring_coredump(struct ctl_table *table, int write,
1089		  void *buffer, size_t *lenp, loff_t *ppos)
1090{
1091	int error = proc_dostring(table, write, buffer, lenp, ppos);
1092	if (!error)
1093		validate_coredump_safety();
1094	return error;
1095}
1096#endif
 
 
1097
1098#ifdef CONFIG_MAGIC_SYSRQ
 
 
 
1099static int sysrq_sysctl_handler(struct ctl_table *table, int write,
1100				void *buffer, size_t *lenp, loff_t *ppos)
 
1101{
1102	int tmp, ret;
1103
1104	tmp = sysrq_mask();
1105
1106	ret = __do_proc_dointvec(&tmp, table, write, buffer,
1107			       lenp, ppos, NULL, NULL);
1108	if (ret || !write)
1109		return ret;
1110
1111	if (write)
1112		sysrq_toggle_support(tmp);
1113
1114	return 0;
1115}
 
1116#endif
1117
1118static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table,
1119		int write, void *buffer, size_t *lenp, loff_t *ppos,
1120		unsigned long convmul, unsigned long convdiv)
1121{
1122	unsigned long *i, *min, *max;
1123	int vleft, first = 1, err = 0;
1124	size_t left;
1125	char *p;
 
1126
1127	if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
1128		*lenp = 0;
1129		return 0;
1130	}
1131
1132	i = (unsigned long *) data;
1133	min = (unsigned long *) table->extra1;
1134	max = (unsigned long *) table->extra2;
1135	vleft = table->maxlen / sizeof(unsigned long);
1136	left = *lenp;
1137
1138	if (write) {
1139		if (proc_first_pos_non_zero_ignore(ppos, table))
1140			goto out;
1141
1142		if (left > PAGE_SIZE - 1)
1143			left = PAGE_SIZE - 1;
1144		p = buffer;
1145	}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1146
1147	for (; left && vleft--; i++, first = 0) {
1148		unsigned long val;
 
 
 
 
 
 
 
 
1149
1150		if (write) {
1151			bool neg;
1152
1153			left -= proc_skip_spaces(&p);
1154			if (!left)
1155				break;
1156
1157			err = proc_get_long(&p, &left, &val, &neg,
1158					     proc_wspace_sep,
1159					     sizeof(proc_wspace_sep), NULL);
1160			if (err)
1161				break;
1162			if (neg)
1163				continue;
1164			val = convmul * val / convdiv;
1165			if ((min && val < *min) || (max && val > *max)) {
1166				err = -EINVAL;
1167				break;
1168			}
1169			*i = val;
1170		} else {
1171			val = convdiv * (*i) / convmul;
1172			if (!first)
1173				proc_put_char(&buffer, &left, '\t');
1174			proc_put_long(&buffer, &left, val, false);
1175		}
1176	}
1177
1178	if (!write && !first && left && !err)
1179		proc_put_char(&buffer, &left, '\n');
1180	if (write && !err)
1181		left -= proc_skip_spaces(&p);
1182	if (write && first)
1183		return err ? : -EINVAL;
1184	*lenp -= left;
1185out:
1186	*ppos += *lenp;
1187	return err;
1188}
1189
1190static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
1191		void *buffer, size_t *lenp, loff_t *ppos, unsigned long convmul,
1192		unsigned long convdiv)
1193{
1194	return __do_proc_doulongvec_minmax(table->data, table, write,
1195			buffer, lenp, ppos, convmul, convdiv);
1196}
1197
1198/**
1199 * proc_doulongvec_minmax - read a vector of long integers with min/max values
1200 * @table: the sysctl table
1201 * @write: %TRUE if this is a write to the sysctl file
1202 * @buffer: the user buffer
1203 * @lenp: the size of the user buffer
1204 * @ppos: file position
1205 *
1206 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1207 * values from/to the user buffer, treated as an ASCII string.
1208 *
1209 * This routine will ensure the values are within the range specified by
1210 * table->extra1 (min) and table->extra2 (max).
1211 *
1212 * Returns 0 on success.
1213 */
1214int proc_doulongvec_minmax(struct ctl_table *table, int write,
1215			   void *buffer, size_t *lenp, loff_t *ppos)
1216{
1217    return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
1218}
1219
1220/**
1221 * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
1222 * @table: the sysctl table
1223 * @write: %TRUE if this is a write to the sysctl file
1224 * @buffer: the user buffer
1225 * @lenp: the size of the user buffer
1226 * @ppos: file position
1227 *
1228 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
1229 * values from/to the user buffer, treated as an ASCII string. The values
1230 * are treated as milliseconds, and converted to jiffies when they are stored.
1231 *
1232 * This routine will ensure the values are within the range specified by
1233 * table->extra1 (min) and table->extra2 (max).
1234 *
1235 * Returns 0 on success.
1236 */
1237int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1238				      void *buffer, size_t *lenp, loff_t *ppos)
1239{
1240    return do_proc_doulongvec_minmax(table, write, buffer,
1241				     lenp, ppos, HZ, 1000l);
1242}
1243
1244
1245static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
1246					 int *valp,
1247					 int write, void *data)
1248{
1249	if (write) {
1250		if (*lvalp > INT_MAX / HZ)
1251			return 1;
1252		*valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
1253	} else {
1254		int val = *valp;
1255		unsigned long lval;
1256		if (val < 0) {
1257			*negp = true;
1258			lval = -(unsigned long)val;
1259		} else {
1260			*negp = false;
1261			lval = (unsigned long)val;
1262		}
1263		*lvalp = lval / HZ;
1264	}
1265	return 0;
1266}
1267
1268static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
1269						int *valp,
1270						int write, void *data)
1271{
1272	if (write) {
1273		if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
1274			return 1;
1275		*valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
1276	} else {
1277		int val = *valp;
1278		unsigned long lval;
1279		if (val < 0) {
1280			*negp = true;
1281			lval = -(unsigned long)val;
1282		} else {
1283			*negp = false;
1284			lval = (unsigned long)val;
1285		}
1286		*lvalp = jiffies_to_clock_t(lval);
1287	}
1288	return 0;
1289}
1290
1291static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
1292					    int *valp,
1293					    int write, void *data)
1294{
1295	if (write) {
1296		unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
1297
1298		if (jif > INT_MAX)
1299			return 1;
1300		*valp = (int)jif;
1301	} else {
1302		int val = *valp;
1303		unsigned long lval;
1304		if (val < 0) {
1305			*negp = true;
1306			lval = -(unsigned long)val;
1307		} else {
1308			*negp = false;
1309			lval = (unsigned long)val;
1310		}
1311		*lvalp = jiffies_to_msecs(lval);
1312	}
1313	return 0;
1314}
1315
1316/**
1317 * proc_dointvec_jiffies - read a vector of integers as seconds
1318 * @table: the sysctl table
1319 * @write: %TRUE if this is a write to the sysctl file
1320 * @buffer: the user buffer
1321 * @lenp: the size of the user buffer
1322 * @ppos: file position
1323 *
1324 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1325 * values from/to the user buffer, treated as an ASCII string. 
1326 * The values read are assumed to be in seconds, and are converted into
1327 * jiffies.
1328 *
1329 * Returns 0 on success.
1330 */
1331int proc_dointvec_jiffies(struct ctl_table *table, int write,
1332			  void *buffer, size_t *lenp, loff_t *ppos)
1333{
1334    return do_proc_dointvec(table,write,buffer,lenp,ppos,
1335		    	    do_proc_dointvec_jiffies_conv,NULL);
1336}
1337
1338/**
1339 * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
1340 * @table: the sysctl table
1341 * @write: %TRUE if this is a write to the sysctl file
1342 * @buffer: the user buffer
1343 * @lenp: the size of the user buffer
1344 * @ppos: pointer to the file position
1345 *
1346 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1347 * values from/to the user buffer, treated as an ASCII string. 
1348 * The values read are assumed to be in 1/USER_HZ seconds, and 
1349 * are converted into jiffies.
1350 *
1351 * Returns 0 on success.
1352 */
1353int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1354				 void *buffer, size_t *lenp, loff_t *ppos)
1355{
1356    return do_proc_dointvec(table,write,buffer,lenp,ppos,
1357		    	    do_proc_dointvec_userhz_jiffies_conv,NULL);
1358}
1359
1360/**
1361 * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
1362 * @table: the sysctl table
1363 * @write: %TRUE if this is a write to the sysctl file
1364 * @buffer: the user buffer
1365 * @lenp: the size of the user buffer
1366 * @ppos: file position
1367 * @ppos: the current position in the file
1368 *
1369 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1370 * values from/to the user buffer, treated as an ASCII string. 
1371 * The values read are assumed to be in 1/1000 seconds, and 
1372 * are converted into jiffies.
1373 *
1374 * Returns 0 on success.
1375 */
1376int proc_dointvec_ms_jiffies(struct ctl_table *table, int write, void *buffer,
1377		size_t *lenp, loff_t *ppos)
1378{
1379	return do_proc_dointvec(table, write, buffer, lenp, ppos,
1380				do_proc_dointvec_ms_jiffies_conv, NULL);
1381}
1382
1383static int proc_do_cad_pid(struct ctl_table *table, int write, void *buffer,
1384		size_t *lenp, loff_t *ppos)
1385{
1386	struct pid *new_pid;
1387	pid_t tmp;
1388	int r;
1389
1390	tmp = pid_vnr(cad_pid);
1391
1392	r = __do_proc_dointvec(&tmp, table, write, buffer,
1393			       lenp, ppos, NULL, NULL);
1394	if (r || !write)
1395		return r;
1396
1397	new_pid = find_get_pid(tmp);
1398	if (!new_pid)
1399		return -ESRCH;
1400
1401	put_pid(xchg(&cad_pid, new_pid));
1402	return 0;
1403}
1404
1405/**
1406 * proc_do_large_bitmap - read/write from/to a large bitmap
1407 * @table: the sysctl table
1408 * @write: %TRUE if this is a write to the sysctl file
1409 * @buffer: the user buffer
1410 * @lenp: the size of the user buffer
1411 * @ppos: file position
1412 *
1413 * The bitmap is stored at table->data and the bitmap length (in bits)
1414 * in table->maxlen.
1415 *
1416 * We use a range comma separated format (e.g. 1,3-4,10-10) so that
1417 * large bitmaps may be represented in a compact manner. Writing into
1418 * the file will clear the bitmap then update it with the given input.
1419 *
1420 * Returns 0 on success.
1421 */
1422int proc_do_large_bitmap(struct ctl_table *table, int write,
1423			 void *buffer, size_t *lenp, loff_t *ppos)
1424{
1425	int err = 0;
1426	bool first = 1;
1427	size_t left = *lenp;
1428	unsigned long bitmap_len = table->maxlen;
1429	unsigned long *bitmap = *(unsigned long **) table->data;
1430	unsigned long *tmp_bitmap = NULL;
1431	char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
1432
1433	if (!bitmap || !bitmap_len || !left || (*ppos && !write)) {
1434		*lenp = 0;
1435		return 0;
1436	}
1437
1438	if (write) {
1439		char *p = buffer;
1440		size_t skipped = 0;
1441
1442		if (left > PAGE_SIZE - 1) {
1443			left = PAGE_SIZE - 1;
1444			/* How much of the buffer we'll skip this pass */
1445			skipped = *lenp - left;
1446		}
1447
1448		tmp_bitmap = bitmap_zalloc(bitmap_len, GFP_KERNEL);
1449		if (!tmp_bitmap)
1450			return -ENOMEM;
1451		proc_skip_char(&p, &left, '\n');
1452		while (!err && left) {
1453			unsigned long val_a, val_b;
1454			bool neg;
1455			size_t saved_left;
1456
1457			/* In case we stop parsing mid-number, we can reset */
1458			saved_left = left;
1459			err = proc_get_long(&p, &left, &val_a, &neg, tr_a,
1460					     sizeof(tr_a), &c);
1461			/*
1462			 * If we consumed the entirety of a truncated buffer or
1463			 * only one char is left (may be a "-"), then stop here,
1464			 * reset, & come back for more.
1465			 */
1466			if ((left <= 1) && skipped) {
1467				left = saved_left;
1468				break;
1469			}
1470
1471			if (err)
1472				break;
1473			if (val_a >= bitmap_len || neg) {
1474				err = -EINVAL;
1475				break;
1476			}
1477
1478			val_b = val_a;
1479			if (left) {
1480				p++;
1481				left--;
1482			}
1483
1484			if (c == '-') {
1485				err = proc_get_long(&p, &left, &val_b,
1486						     &neg, tr_b, sizeof(tr_b),
1487						     &c);
1488				/*
1489				 * If we consumed all of a truncated buffer or
1490				 * then stop here, reset, & come back for more.
1491				 */
1492				if (!left && skipped) {
1493					left = saved_left;
1494					break;
1495				}
1496
1497				if (err)
1498					break;
1499				if (val_b >= bitmap_len || neg ||
1500				    val_a > val_b) {
1501					err = -EINVAL;
1502					break;
1503				}
1504				if (left) {
1505					p++;
1506					left--;
1507				}
1508			}
1509
1510			bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
1511			first = 0;
1512			proc_skip_char(&p, &left, '\n');
1513		}
1514		left += skipped;
1515	} else {
1516		unsigned long bit_a, bit_b = 0;
1517
1518		while (left) {
1519			bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
1520			if (bit_a >= bitmap_len)
1521				break;
1522			bit_b = find_next_zero_bit(bitmap, bitmap_len,
1523						   bit_a + 1) - 1;
1524
1525			if (!first)
1526				proc_put_char(&buffer, &left, ',');
1527			proc_put_long(&buffer, &left, bit_a, false);
1528			if (bit_a != bit_b) {
1529				proc_put_char(&buffer, &left, '-');
1530				proc_put_long(&buffer, &left, bit_b, false);
1531			}
1532
1533			first = 0; bit_b++;
1534		}
1535		proc_put_char(&buffer, &left, '\n');
1536	}
1537
1538	if (!err) {
1539		if (write) {
1540			if (*ppos)
1541				bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
1542			else
1543				bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
1544		}
1545		*lenp -= left;
1546		*ppos += *lenp;
1547	}
1548
1549	bitmap_free(tmp_bitmap);
1550	return err;
1551}
1552
1553#else /* CONFIG_PROC_SYSCTL */
1554
1555int proc_dostring(struct ctl_table *table, int write,
1556		  void *buffer, size_t *lenp, loff_t *ppos)
1557{
1558	return -ENOSYS;
1559}
1560
1561int proc_dointvec(struct ctl_table *table, int write,
1562		  void *buffer, size_t *lenp, loff_t *ppos)
1563{
1564	return -ENOSYS;
1565}
1566
1567int proc_douintvec(struct ctl_table *table, int write,
1568		  void *buffer, size_t *lenp, loff_t *ppos)
1569{
1570	return -ENOSYS;
1571}
1572
1573int proc_dointvec_minmax(struct ctl_table *table, int write,
1574		    void *buffer, size_t *lenp, loff_t *ppos)
1575{
1576	return -ENOSYS;
1577}
1578
1579int proc_douintvec_minmax(struct ctl_table *table, int write,
1580			  void *buffer, size_t *lenp, loff_t *ppos)
1581{
1582	return -ENOSYS;
1583}
1584
1585int proc_dointvec_jiffies(struct ctl_table *table, int write,
1586		    void *buffer, size_t *lenp, loff_t *ppos)
1587{
1588	return -ENOSYS;
1589}
1590
1591int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
1592		    void *buffer, size_t *lenp, loff_t *ppos)
1593{
1594	return -ENOSYS;
1595}
1596
1597int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
1598			     void *buffer, size_t *lenp, loff_t *ppos)
1599{
1600	return -ENOSYS;
1601}
1602
1603int proc_doulongvec_minmax(struct ctl_table *table, int write,
1604		    void *buffer, size_t *lenp, loff_t *ppos)
1605{
1606	return -ENOSYS;
1607}
1608
1609int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
1610				      void *buffer, size_t *lenp, loff_t *ppos)
1611{
1612	return -ENOSYS;
1613}
1614
1615int proc_do_large_bitmap(struct ctl_table *table, int write,
1616			 void *buffer, size_t *lenp, loff_t *ppos)
1617{
1618	return -ENOSYS;
1619}
1620
1621#endif /* CONFIG_PROC_SYSCTL */
1622
1623#if defined(CONFIG_SYSCTL)
1624int proc_do_static_key(struct ctl_table *table, int write,
1625		       void *buffer, size_t *lenp, loff_t *ppos)
1626{
1627	struct static_key *key = (struct static_key *)table->data;
1628	static DEFINE_MUTEX(static_key_mutex);
1629	int val, ret;
1630	struct ctl_table tmp = {
1631		.data   = &val,
1632		.maxlen = sizeof(val),
1633		.mode   = table->mode,
1634		.extra1 = SYSCTL_ZERO,
1635		.extra2 = SYSCTL_ONE,
1636	};
1637
1638	if (write && !capable(CAP_SYS_ADMIN))
1639		return -EPERM;
1640
1641	mutex_lock(&static_key_mutex);
1642	val = static_key_enabled(key);
1643	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
1644	if (write && !ret) {
1645		if (val)
1646			static_key_enable(key);
1647		else
1648			static_key_disable(key);
1649	}
1650	mutex_unlock(&static_key_mutex);
1651	return ret;
1652}
1653
1654static struct ctl_table kern_table[] = {
1655	{
1656		.procname	= "sched_child_runs_first",
1657		.data		= &sysctl_sched_child_runs_first,
1658		.maxlen		= sizeof(unsigned int),
1659		.mode		= 0644,
1660		.proc_handler	= proc_dointvec,
1661	},
1662#ifdef CONFIG_SCHED_DEBUG
1663	{
1664		.procname	= "sched_min_granularity_ns",
1665		.data		= &sysctl_sched_min_granularity,
1666		.maxlen		= sizeof(unsigned int),
1667		.mode		= 0644,
1668		.proc_handler	= sched_proc_update_handler,
1669		.extra1		= &min_sched_granularity_ns,
1670		.extra2		= &max_sched_granularity_ns,
1671	},
1672	{
1673		.procname	= "sched_latency_ns",
1674		.data		= &sysctl_sched_latency,
1675		.maxlen		= sizeof(unsigned int),
1676		.mode		= 0644,
1677		.proc_handler	= sched_proc_update_handler,
1678		.extra1		= &min_sched_granularity_ns,
1679		.extra2		= &max_sched_granularity_ns,
1680	},
1681	{
1682		.procname	= "sched_wakeup_granularity_ns",
1683		.data		= &sysctl_sched_wakeup_granularity,
1684		.maxlen		= sizeof(unsigned int),
1685		.mode		= 0644,
1686		.proc_handler	= sched_proc_update_handler,
1687		.extra1		= &min_wakeup_granularity_ns,
1688		.extra2		= &max_wakeup_granularity_ns,
1689	},
1690#ifdef CONFIG_SMP
1691	{
1692		.procname	= "sched_tunable_scaling",
1693		.data		= &sysctl_sched_tunable_scaling,
1694		.maxlen		= sizeof(enum sched_tunable_scaling),
1695		.mode		= 0644,
1696		.proc_handler	= sched_proc_update_handler,
1697		.extra1		= &min_sched_tunable_scaling,
1698		.extra2		= &max_sched_tunable_scaling,
1699	},
1700	{
1701		.procname	= "sched_migration_cost_ns",
1702		.data		= &sysctl_sched_migration_cost,
1703		.maxlen		= sizeof(unsigned int),
1704		.mode		= 0644,
1705		.proc_handler	= proc_dointvec,
1706	},
1707	{
1708		.procname	= "sched_nr_migrate",
1709		.data		= &sysctl_sched_nr_migrate,
1710		.maxlen		= sizeof(unsigned int),
1711		.mode		= 0644,
1712		.proc_handler	= proc_dointvec,
1713	},
1714#ifdef CONFIG_SCHEDSTATS
1715	{
1716		.procname	= "sched_schedstats",
1717		.data		= NULL,
1718		.maxlen		= sizeof(unsigned int),
1719		.mode		= 0644,
1720		.proc_handler	= sysctl_schedstats,
1721		.extra1		= SYSCTL_ZERO,
1722		.extra2		= SYSCTL_ONE,
1723	},
1724#endif /* CONFIG_SCHEDSTATS */
1725#endif /* CONFIG_SMP */
1726#ifdef CONFIG_NUMA_BALANCING
1727	{
1728		.procname	= "numa_balancing_scan_delay_ms",
1729		.data		= &sysctl_numa_balancing_scan_delay,
1730		.maxlen		= sizeof(unsigned int),
1731		.mode		= 0644,
1732		.proc_handler	= proc_dointvec,
1733	},
1734	{
1735		.procname	= "numa_balancing_scan_period_min_ms",
1736		.data		= &sysctl_numa_balancing_scan_period_min,
1737		.maxlen		= sizeof(unsigned int),
1738		.mode		= 0644,
1739		.proc_handler	= proc_dointvec,
1740	},
1741	{
1742		.procname	= "numa_balancing_scan_period_max_ms",
1743		.data		= &sysctl_numa_balancing_scan_period_max,
1744		.maxlen		= sizeof(unsigned int),
1745		.mode		= 0644,
1746		.proc_handler	= proc_dointvec,
1747	},
1748	{
1749		.procname	= "numa_balancing_scan_size_mb",
1750		.data		= &sysctl_numa_balancing_scan_size,
1751		.maxlen		= sizeof(unsigned int),
1752		.mode		= 0644,
1753		.proc_handler	= proc_dointvec_minmax,
1754		.extra1		= SYSCTL_ONE,
1755	},
1756	{
1757		.procname	= "numa_balancing",
1758		.data		= NULL, /* filled in by handler */
1759		.maxlen		= sizeof(unsigned int),
1760		.mode		= 0644,
1761		.proc_handler	= sysctl_numa_balancing,
1762		.extra1		= SYSCTL_ZERO,
1763		.extra2		= SYSCTL_ONE,
1764	},
1765#endif /* CONFIG_NUMA_BALANCING */
1766#endif /* CONFIG_SCHED_DEBUG */
1767	{
1768		.procname	= "sched_rt_period_us",
1769		.data		= &sysctl_sched_rt_period,
1770		.maxlen		= sizeof(unsigned int),
1771		.mode		= 0644,
1772		.proc_handler	= sched_rt_handler,
1773	},
1774	{
1775		.procname	= "sched_rt_runtime_us",
1776		.data		= &sysctl_sched_rt_runtime,
1777		.maxlen		= sizeof(int),
1778		.mode		= 0644,
1779		.proc_handler	= sched_rt_handler,
1780	},
1781	{
1782		.procname	= "sched_deadline_period_max_us",
1783		.data		= &sysctl_sched_dl_period_max,
1784		.maxlen		= sizeof(unsigned int),
1785		.mode		= 0644,
1786		.proc_handler	= proc_dointvec,
1787	},
1788	{
1789		.procname	= "sched_deadline_period_min_us",
1790		.data		= &sysctl_sched_dl_period_min,
1791		.maxlen		= sizeof(unsigned int),
1792		.mode		= 0644,
1793		.proc_handler	= proc_dointvec,
1794	},
1795	{
1796		.procname	= "sched_rr_timeslice_ms",
1797		.data		= &sysctl_sched_rr_timeslice,
1798		.maxlen		= sizeof(int),
1799		.mode		= 0644,
1800		.proc_handler	= sched_rr_handler,
1801	},
1802#ifdef CONFIG_UCLAMP_TASK
1803	{
1804		.procname	= "sched_util_clamp_min",
1805		.data		= &sysctl_sched_uclamp_util_min,
1806		.maxlen		= sizeof(unsigned int),
1807		.mode		= 0644,
1808		.proc_handler	= sysctl_sched_uclamp_handler,
1809	},
1810	{
1811		.procname	= "sched_util_clamp_max",
1812		.data		= &sysctl_sched_uclamp_util_max,
1813		.maxlen		= sizeof(unsigned int),
1814		.mode		= 0644,
1815		.proc_handler	= sysctl_sched_uclamp_handler,
1816	},
1817	{
1818		.procname	= "sched_util_clamp_min_rt_default",
1819		.data		= &sysctl_sched_uclamp_util_min_rt_default,
1820		.maxlen		= sizeof(unsigned int),
1821		.mode		= 0644,
1822		.proc_handler	= sysctl_sched_uclamp_handler,
1823	},
1824#endif
1825#ifdef CONFIG_SCHED_AUTOGROUP
1826	{
1827		.procname	= "sched_autogroup_enabled",
1828		.data		= &sysctl_sched_autogroup_enabled,
1829		.maxlen		= sizeof(unsigned int),
1830		.mode		= 0644,
1831		.proc_handler	= proc_dointvec_minmax,
1832		.extra1		= SYSCTL_ZERO,
1833		.extra2		= SYSCTL_ONE,
1834	},
1835#endif
1836#ifdef CONFIG_CFS_BANDWIDTH
1837	{
1838		.procname	= "sched_cfs_bandwidth_slice_us",
1839		.data		= &sysctl_sched_cfs_bandwidth_slice,
1840		.maxlen		= sizeof(unsigned int),
1841		.mode		= 0644,
1842		.proc_handler	= proc_dointvec_minmax,
1843		.extra1		= SYSCTL_ONE,
1844	},
1845#endif
1846#if defined(CONFIG_ENERGY_MODEL) && defined(CONFIG_CPU_FREQ_GOV_SCHEDUTIL)
1847	{
1848		.procname	= "sched_energy_aware",
1849		.data		= &sysctl_sched_energy_aware,
1850		.maxlen		= sizeof(unsigned int),
1851		.mode		= 0644,
1852		.proc_handler	= sched_energy_aware_handler,
1853		.extra1		= SYSCTL_ZERO,
1854		.extra2		= SYSCTL_ONE,
1855	},
1856#endif
1857#ifdef CONFIG_PROVE_LOCKING
1858	{
1859		.procname	= "prove_locking",
1860		.data		= &prove_locking,
1861		.maxlen		= sizeof(int),
1862		.mode		= 0644,
1863		.proc_handler	= proc_dointvec,
1864	},
1865#endif
1866#ifdef CONFIG_LOCK_STAT
1867	{
1868		.procname	= "lock_stat",
1869		.data		= &lock_stat,
1870		.maxlen		= sizeof(int),
1871		.mode		= 0644,
1872		.proc_handler	= proc_dointvec,
1873	},
1874#endif
1875	{
1876		.procname	= "panic",
1877		.data		= &panic_timeout,
1878		.maxlen		= sizeof(int),
1879		.mode		= 0644,
1880		.proc_handler	= proc_dointvec,
1881	},
1882#ifdef CONFIG_COREDUMP
1883	{
1884		.procname	= "core_uses_pid",
1885		.data		= &core_uses_pid,
1886		.maxlen		= sizeof(int),
1887		.mode		= 0644,
1888		.proc_handler	= proc_dointvec,
1889	},
1890	{
1891		.procname	= "core_pattern",
1892		.data		= core_pattern,
1893		.maxlen		= CORENAME_MAX_SIZE,
1894		.mode		= 0644,
1895		.proc_handler	= proc_dostring_coredump,
1896	},
1897	{
1898		.procname	= "core_pipe_limit",
1899		.data		= &core_pipe_limit,
1900		.maxlen		= sizeof(unsigned int),
1901		.mode		= 0644,
1902		.proc_handler	= proc_dointvec,
1903	},
1904#endif
1905#ifdef CONFIG_PROC_SYSCTL
1906	{
1907		.procname	= "tainted",
1908		.maxlen 	= sizeof(long),
1909		.mode		= 0644,
1910		.proc_handler	= proc_taint,
1911	},
1912	{
1913		.procname	= "sysctl_writes_strict",
1914		.data		= &sysctl_writes_strict,
1915		.maxlen		= sizeof(int),
1916		.mode		= 0644,
1917		.proc_handler	= proc_dointvec_minmax,
1918		.extra1		= &neg_one,
1919		.extra2		= SYSCTL_ONE,
1920	},
1921#endif
1922#ifdef CONFIG_LATENCYTOP
1923	{
1924		.procname	= "latencytop",
1925		.data		= &latencytop_enabled,
1926		.maxlen		= sizeof(int),
1927		.mode		= 0644,
1928		.proc_handler	= sysctl_latencytop,
1929	},
1930#endif
1931#ifdef CONFIG_BLK_DEV_INITRD
1932	{
1933		.procname	= "real-root-dev",
1934		.data		= &real_root_dev,
1935		.maxlen		= sizeof(int),
1936		.mode		= 0644,
1937		.proc_handler	= proc_dointvec,
1938	},
1939#endif
1940	{
1941		.procname	= "print-fatal-signals",
1942		.data		= &print_fatal_signals,
1943		.maxlen		= sizeof(int),
1944		.mode		= 0644,
1945		.proc_handler	= proc_dointvec,
1946	},
1947#ifdef CONFIG_SPARC
1948	{
1949		.procname	= "reboot-cmd",
1950		.data		= reboot_command,
1951		.maxlen		= 256,
1952		.mode		= 0644,
1953		.proc_handler	= proc_dostring,
1954	},
1955	{
1956		.procname	= "stop-a",
1957		.data		= &stop_a_enabled,
1958		.maxlen		= sizeof (int),
1959		.mode		= 0644,
1960		.proc_handler	= proc_dointvec,
1961	},
1962	{
1963		.procname	= "scons-poweroff",
1964		.data		= &scons_pwroff,
1965		.maxlen		= sizeof (int),
1966		.mode		= 0644,
1967		.proc_handler	= proc_dointvec,
1968	},
1969#endif
1970#ifdef CONFIG_SPARC64
1971	{
1972		.procname	= "tsb-ratio",
1973		.data		= &sysctl_tsb_ratio,
1974		.maxlen		= sizeof (int),
1975		.mode		= 0644,
1976		.proc_handler	= proc_dointvec,
1977	},
1978#endif
1979#ifdef CONFIG_PARISC
1980	{
1981		.procname	= "soft-power",
1982		.data		= &pwrsw_enabled,
1983		.maxlen		= sizeof (int),
1984		.mode		= 0644,
1985		.proc_handler	= proc_dointvec,
1986	},
1987#endif
1988#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
1989	{
1990		.procname	= "unaligned-trap",
1991		.data		= &unaligned_enabled,
1992		.maxlen		= sizeof (int),
1993		.mode		= 0644,
1994		.proc_handler	= proc_dointvec,
1995	},
1996#endif
1997	{
1998		.procname	= "ctrl-alt-del",
1999		.data		= &C_A_D,
2000		.maxlen		= sizeof(int),
2001		.mode		= 0644,
2002		.proc_handler	= proc_dointvec,
2003	},
2004#ifdef CONFIG_FUNCTION_TRACER
2005	{
2006		.procname	= "ftrace_enabled",
2007		.data		= &ftrace_enabled,
2008		.maxlen		= sizeof(int),
2009		.mode		= 0644,
2010		.proc_handler	= ftrace_enable_sysctl,
2011	},
2012#endif
2013#ifdef CONFIG_STACK_TRACER
2014	{
2015		.procname	= "stack_tracer_enabled",
2016		.data		= &stack_tracer_enabled,
2017		.maxlen		= sizeof(int),
2018		.mode		= 0644,
2019		.proc_handler	= stack_trace_sysctl,
2020	},
2021#endif
2022#ifdef CONFIG_TRACING
2023	{
2024		.procname	= "ftrace_dump_on_oops",
2025		.data		= &ftrace_dump_on_oops,
2026		.maxlen		= sizeof(int),
2027		.mode		= 0644,
2028		.proc_handler	= proc_dointvec,
2029	},
2030	{
2031		.procname	= "traceoff_on_warning",
2032		.data		= &__disable_trace_on_warning,
2033		.maxlen		= sizeof(__disable_trace_on_warning),
2034		.mode		= 0644,
2035		.proc_handler	= proc_dointvec,
2036	},
2037	{
2038		.procname	= "tracepoint_printk",
2039		.data		= &tracepoint_printk,
2040		.maxlen		= sizeof(tracepoint_printk),
2041		.mode		= 0644,
2042		.proc_handler	= tracepoint_printk_sysctl,
2043	},
2044#endif
2045#ifdef CONFIG_KEXEC_CORE
2046	{
2047		.procname	= "kexec_load_disabled",
2048		.data		= &kexec_load_disabled,
2049		.maxlen		= sizeof(int),
2050		.mode		= 0644,
2051		/* only handle a transition from default "0" to "1" */
2052		.proc_handler	= proc_dointvec_minmax,
2053		.extra1		= SYSCTL_ONE,
2054		.extra2		= SYSCTL_ONE,
2055	},
2056#endif
2057#ifdef CONFIG_MODULES
2058	{
2059		.procname	= "modprobe",
2060		.data		= &modprobe_path,
2061		.maxlen		= KMOD_PATH_LEN,
2062		.mode		= 0644,
2063		.proc_handler	= proc_dostring,
2064	},
2065	{
2066		.procname	= "modules_disabled",
2067		.data		= &modules_disabled,
2068		.maxlen		= sizeof(int),
2069		.mode		= 0644,
2070		/* only handle a transition from default "0" to "1" */
2071		.proc_handler	= proc_dointvec_minmax,
2072		.extra1		= SYSCTL_ONE,
2073		.extra2		= SYSCTL_ONE,
2074	},
2075#endif
2076#ifdef CONFIG_UEVENT_HELPER
2077	{
2078		.procname	= "hotplug",
2079		.data		= &uevent_helper,
2080		.maxlen		= UEVENT_HELPER_PATH_LEN,
2081		.mode		= 0644,
2082		.proc_handler	= proc_dostring,
2083	},
2084#endif
2085#ifdef CONFIG_CHR_DEV_SG
2086	{
2087		.procname	= "sg-big-buff",
2088		.data		= &sg_big_buff,
2089		.maxlen		= sizeof (int),
2090		.mode		= 0444,
2091		.proc_handler	= proc_dointvec,
2092	},
2093#endif
2094#ifdef CONFIG_BSD_PROCESS_ACCT
2095	{
2096		.procname	= "acct",
2097		.data		= &acct_parm,
2098		.maxlen		= 3*sizeof(int),
2099		.mode		= 0644,
2100		.proc_handler	= proc_dointvec,
2101	},
2102#endif
2103#ifdef CONFIG_MAGIC_SYSRQ
2104	{
2105		.procname	= "sysrq",
2106		.data		= NULL,
2107		.maxlen		= sizeof (int),
2108		.mode		= 0644,
2109		.proc_handler	= sysrq_sysctl_handler,
2110	},
2111#endif
2112#ifdef CONFIG_PROC_SYSCTL
2113	{
2114		.procname	= "cad_pid",
2115		.data		= NULL,
2116		.maxlen		= sizeof (int),
2117		.mode		= 0600,
2118		.proc_handler	= proc_do_cad_pid,
2119	},
2120#endif
2121	{
2122		.procname	= "threads-max",
2123		.data		= NULL,
2124		.maxlen		= sizeof(int),
2125		.mode		= 0644,
2126		.proc_handler	= sysctl_max_threads,
2127	},
2128	{
2129		.procname	= "random",
2130		.mode		= 0555,
2131		.child		= random_table,
2132	},
2133	{
2134		.procname	= "usermodehelper",
2135		.mode		= 0555,
2136		.child		= usermodehelper_table,
2137	},
2138#ifdef CONFIG_FW_LOADER_USER_HELPER
2139	{
2140		.procname	= "firmware_config",
2141		.mode		= 0555,
2142		.child		= firmware_config_table,
2143	},
2144#endif
2145	{
2146		.procname	= "overflowuid",
2147		.data		= &overflowuid,
2148		.maxlen		= sizeof(int),
2149		.mode		= 0644,
2150		.proc_handler	= proc_dointvec_minmax,
2151		.extra1		= &minolduid,
2152		.extra2		= &maxolduid,
2153	},
2154	{
2155		.procname	= "overflowgid",
2156		.data		= &overflowgid,
2157		.maxlen		= sizeof(int),
2158		.mode		= 0644,
2159		.proc_handler	= proc_dointvec_minmax,
2160		.extra1		= &minolduid,
2161		.extra2		= &maxolduid,
2162	},
2163#ifdef CONFIG_S390
 
2164	{
2165		.procname	= "userprocess_debug",
2166		.data		= &show_unhandled_signals,
2167		.maxlen		= sizeof(int),
2168		.mode		= 0644,
2169		.proc_handler	= proc_dointvec,
2170	},
2171#endif
2172#ifdef CONFIG_SMP
2173	{
2174		.procname	= "oops_all_cpu_backtrace",
2175		.data		= &sysctl_oops_all_cpu_backtrace,
2176		.maxlen		= sizeof(int),
2177		.mode		= 0644,
2178		.proc_handler	= proc_dointvec_minmax,
2179		.extra1		= SYSCTL_ZERO,
2180		.extra2		= SYSCTL_ONE,
2181	},
2182#endif /* CONFIG_SMP */
2183	{
2184		.procname	= "pid_max",
2185		.data		= &pid_max,
2186		.maxlen		= sizeof (int),
2187		.mode		= 0644,
2188		.proc_handler	= proc_dointvec_minmax,
2189		.extra1		= &pid_max_min,
2190		.extra2		= &pid_max_max,
2191	},
2192	{
2193		.procname	= "panic_on_oops",
2194		.data		= &panic_on_oops,
2195		.maxlen		= sizeof(int),
2196		.mode		= 0644,
2197		.proc_handler	= proc_dointvec,
2198	},
2199	{
2200		.procname	= "panic_print",
2201		.data		= &panic_print,
2202		.maxlen		= sizeof(unsigned long),
2203		.mode		= 0644,
2204		.proc_handler	= proc_doulongvec_minmax,
2205	},
2206#if defined CONFIG_PRINTK
2207	{
2208		.procname	= "printk",
2209		.data		= &console_loglevel,
2210		.maxlen		= 4*sizeof(int),
2211		.mode		= 0644,
2212		.proc_handler	= proc_dointvec,
2213	},
2214	{
2215		.procname	= "printk_ratelimit",
2216		.data		= &printk_ratelimit_state.interval,
2217		.maxlen		= sizeof(int),
2218		.mode		= 0644,
2219		.proc_handler	= proc_dointvec_jiffies,
2220	},
2221	{
2222		.procname	= "printk_ratelimit_burst",
2223		.data		= &printk_ratelimit_state.burst,
2224		.maxlen		= sizeof(int),
2225		.mode		= 0644,
2226		.proc_handler	= proc_dointvec,
2227	},
2228	{
2229		.procname	= "printk_delay",
2230		.data		= &printk_delay_msec,
2231		.maxlen		= sizeof(int),
2232		.mode		= 0644,
2233		.proc_handler	= proc_dointvec_minmax,
2234		.extra1		= SYSCTL_ZERO,
2235		.extra2		= &ten_thousand,
2236	},
2237	{
2238		.procname	= "printk_devkmsg",
2239		.data		= devkmsg_log_str,
2240		.maxlen		= DEVKMSG_STR_MAX_SIZE,
2241		.mode		= 0644,
2242		.proc_handler	= devkmsg_sysctl_set_loglvl,
2243	},
2244	{
2245		.procname	= "dmesg_restrict",
2246		.data		= &dmesg_restrict,
2247		.maxlen		= sizeof(int),
2248		.mode		= 0644,
2249		.proc_handler	= proc_dointvec_minmax_sysadmin,
2250		.extra1		= SYSCTL_ZERO,
2251		.extra2		= SYSCTL_ONE,
2252	},
2253	{
2254		.procname	= "kptr_restrict",
2255		.data		= &kptr_restrict,
2256		.maxlen		= sizeof(int),
2257		.mode		= 0644,
2258		.proc_handler	= proc_dointvec_minmax_sysadmin,
2259		.extra1		= SYSCTL_ZERO,
2260		.extra2		= &two,
2261	},
2262#endif
2263	{
2264		.procname	= "ngroups_max",
2265		.data		= &ngroups_max,
2266		.maxlen		= sizeof (int),
2267		.mode		= 0444,
2268		.proc_handler	= proc_dointvec,
2269	},
2270	{
2271		.procname	= "cap_last_cap",
2272		.data		= (void *)&cap_last_cap,
2273		.maxlen		= sizeof(int),
2274		.mode		= 0444,
2275		.proc_handler	= proc_dointvec,
2276	},
2277#if defined(CONFIG_LOCKUP_DETECTOR)
2278	{
2279		.procname       = "watchdog",
2280		.data		= &watchdog_user_enabled,
2281		.maxlen		= sizeof(int),
2282		.mode		= 0644,
2283		.proc_handler   = proc_watchdog,
2284		.extra1		= SYSCTL_ZERO,
2285		.extra2		= SYSCTL_ONE,
2286	},
2287	{
2288		.procname	= "watchdog_thresh",
2289		.data		= &watchdog_thresh,
2290		.maxlen		= sizeof(int),
2291		.mode		= 0644,
2292		.proc_handler	= proc_watchdog_thresh,
2293		.extra1		= SYSCTL_ZERO,
2294		.extra2		= &sixty,
2295	},
2296	{
2297		.procname       = "nmi_watchdog",
2298		.data		= &nmi_watchdog_user_enabled,
2299		.maxlen		= sizeof(int),
2300		.mode		= NMI_WATCHDOG_SYSCTL_PERM,
2301		.proc_handler   = proc_nmi_watchdog,
2302		.extra1		= SYSCTL_ZERO,
2303		.extra2		= SYSCTL_ONE,
2304	},
2305	{
2306		.procname	= "watchdog_cpumask",
2307		.data		= &watchdog_cpumask_bits,
2308		.maxlen		= NR_CPUS,
2309		.mode		= 0644,
2310		.proc_handler	= proc_watchdog_cpumask,
2311	},
2312#ifdef CONFIG_SOFTLOCKUP_DETECTOR
2313	{
2314		.procname       = "soft_watchdog",
2315		.data		= &soft_watchdog_user_enabled,
2316		.maxlen		= sizeof(int),
2317		.mode		= 0644,
2318		.proc_handler   = proc_soft_watchdog,
2319		.extra1		= SYSCTL_ZERO,
2320		.extra2		= SYSCTL_ONE,
2321	},
2322	{
2323		.procname	= "softlockup_panic",
2324		.data		= &softlockup_panic,
2325		.maxlen		= sizeof(int),
2326		.mode		= 0644,
2327		.proc_handler	= proc_dointvec_minmax,
2328		.extra1		= SYSCTL_ZERO,
2329		.extra2		= SYSCTL_ONE,
2330	},
2331#ifdef CONFIG_SMP
2332	{
2333		.procname	= "softlockup_all_cpu_backtrace",
2334		.data		= &sysctl_softlockup_all_cpu_backtrace,
2335		.maxlen		= sizeof(int),
2336		.mode		= 0644,
2337		.proc_handler	= proc_dointvec_minmax,
2338		.extra1		= SYSCTL_ZERO,
2339		.extra2		= SYSCTL_ONE,
2340	},
2341#endif /* CONFIG_SMP */
2342#endif
2343#ifdef CONFIG_HARDLOCKUP_DETECTOR
2344	{
2345		.procname	= "hardlockup_panic",
2346		.data		= &hardlockup_panic,
2347		.maxlen		= sizeof(int),
2348		.mode		= 0644,
2349		.proc_handler	= proc_dointvec_minmax,
2350		.extra1		= SYSCTL_ZERO,
2351		.extra2		= SYSCTL_ONE,
2352	},
2353#ifdef CONFIG_SMP
2354	{
2355		.procname	= "hardlockup_all_cpu_backtrace",
2356		.data		= &sysctl_hardlockup_all_cpu_backtrace,
2357		.maxlen		= sizeof(int),
2358		.mode		= 0644,
2359		.proc_handler	= proc_dointvec_minmax,
2360		.extra1		= SYSCTL_ZERO,
2361		.extra2		= SYSCTL_ONE,
2362	},
2363#endif /* CONFIG_SMP */
2364#endif
2365#endif
2366
2367#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
2368	{
2369		.procname       = "unknown_nmi_panic",
2370		.data           = &unknown_nmi_panic,
2371		.maxlen         = sizeof (int),
2372		.mode           = 0644,
2373		.proc_handler   = proc_dointvec,
2374	},
2375#endif
2376
2377#if (defined(CONFIG_X86_32) || defined(CONFIG_PARISC)) && \
2378	defined(CONFIG_DEBUG_STACKOVERFLOW)
2379	{
2380		.procname	= "panic_on_stackoverflow",
2381		.data		= &sysctl_panic_on_stackoverflow,
2382		.maxlen		= sizeof(int),
2383		.mode		= 0644,
2384		.proc_handler	= proc_dointvec,
2385	},
2386#endif
2387#if defined(CONFIG_X86)
2388	{
2389		.procname	= "panic_on_unrecovered_nmi",
2390		.data		= &panic_on_unrecovered_nmi,
2391		.maxlen		= sizeof(int),
2392		.mode		= 0644,
2393		.proc_handler	= proc_dointvec,
2394	},
 
2395	{
2396		.procname	= "panic_on_io_nmi",
2397		.data		= &panic_on_io_nmi,
2398		.maxlen		= sizeof(int),
2399		.mode		= 0644,
2400		.proc_handler	= proc_dointvec,
2401	},
 
2402	{
2403		.procname	= "bootloader_type",
2404		.data		= &bootloader_type,
2405		.maxlen		= sizeof (int),
2406		.mode		= 0444,
2407		.proc_handler	= proc_dointvec,
2408	},
2409	{
2410		.procname	= "bootloader_version",
2411		.data		= &bootloader_version,
2412		.maxlen		= sizeof (int),
2413		.mode		= 0444,
2414		.proc_handler	= proc_dointvec,
2415	},
2416	{
2417		.procname	= "io_delay_type",
2418		.data		= &io_delay_type,
2419		.maxlen		= sizeof(int),
2420		.mode		= 0644,
2421		.proc_handler	= proc_dointvec,
2422	},
2423#endif
2424#if defined(CONFIG_MMU)
2425	{
2426		.procname	= "randomize_va_space",
2427		.data		= &randomize_va_space,
2428		.maxlen		= sizeof(int),
2429		.mode		= 0644,
2430		.proc_handler	= proc_dointvec,
2431	},
2432#endif
2433#if defined(CONFIG_S390) && defined(CONFIG_SMP)
2434	{
2435		.procname	= "spin_retry",
2436		.data		= &spin_retry,
2437		.maxlen		= sizeof (int),
2438		.mode		= 0644,
2439		.proc_handler	= proc_dointvec,
2440	},
2441#endif
2442#if	defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
2443	{
2444		.procname	= "acpi_video_flags",
2445		.data		= &acpi_realmode_flags,
2446		.maxlen		= sizeof (unsigned long),
2447		.mode		= 0644,
2448		.proc_handler	= proc_doulongvec_minmax,
2449	},
2450#endif
2451#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
2452	{
2453		.procname	= "ignore-unaligned-usertrap",
2454		.data		= &no_unaligned_warning,
2455		.maxlen		= sizeof (int),
2456		.mode		= 0644,
2457		.proc_handler	= proc_dointvec,
2458	},
2459#endif
2460#ifdef CONFIG_IA64
2461	{
2462		.procname	= "unaligned-dump-stack",
2463		.data		= &unaligned_dump_stack,
2464		.maxlen		= sizeof (int),
2465		.mode		= 0644,
2466		.proc_handler	= proc_dointvec,
2467	},
2468#endif
2469#ifdef CONFIG_DETECT_HUNG_TASK
2470#ifdef CONFIG_SMP
2471	{
2472		.procname	= "hung_task_all_cpu_backtrace",
2473		.data		= &sysctl_hung_task_all_cpu_backtrace,
2474		.maxlen		= sizeof(int),
2475		.mode		= 0644,
2476		.proc_handler	= proc_dointvec_minmax,
2477		.extra1		= SYSCTL_ZERO,
2478		.extra2		= SYSCTL_ONE,
2479	},
2480#endif /* CONFIG_SMP */
2481	{
2482		.procname	= "hung_task_panic",
2483		.data		= &sysctl_hung_task_panic,
2484		.maxlen		= sizeof(int),
2485		.mode		= 0644,
2486		.proc_handler	= proc_dointvec_minmax,
2487		.extra1		= SYSCTL_ZERO,
2488		.extra2		= SYSCTL_ONE,
2489	},
2490	{
2491		.procname	= "hung_task_check_count",
2492		.data		= &sysctl_hung_task_check_count,
2493		.maxlen		= sizeof(int),
2494		.mode		= 0644,
2495		.proc_handler	= proc_dointvec_minmax,
2496		.extra1		= SYSCTL_ZERO,
2497	},
2498	{
2499		.procname	= "hung_task_timeout_secs",
2500		.data		= &sysctl_hung_task_timeout_secs,
2501		.maxlen		= sizeof(unsigned long),
2502		.mode		= 0644,
2503		.proc_handler	= proc_dohung_task_timeout_secs,
2504		.extra2		= &hung_task_timeout_max,
2505	},
2506	{
2507		.procname	= "hung_task_check_interval_secs",
2508		.data		= &sysctl_hung_task_check_interval_secs,
2509		.maxlen		= sizeof(unsigned long),
2510		.mode		= 0644,
2511		.proc_handler	= proc_dohung_task_timeout_secs,
2512		.extra2		= &hung_task_timeout_max,
2513	},
2514	{
2515		.procname	= "hung_task_warnings",
2516		.data		= &sysctl_hung_task_warnings,
2517		.maxlen		= sizeof(int),
2518		.mode		= 0644,
2519		.proc_handler	= proc_dointvec_minmax,
2520		.extra1		= &neg_one,
2521	},
2522#endif
2523#ifdef CONFIG_RT_MUTEXES
2524	{
2525		.procname	= "max_lock_depth",
2526		.data		= &max_lock_depth,
2527		.maxlen		= sizeof(int),
2528		.mode		= 0644,
2529		.proc_handler	= proc_dointvec,
2530	},
2531#endif
2532	{
2533		.procname	= "poweroff_cmd",
2534		.data		= &poweroff_cmd,
2535		.maxlen		= POWEROFF_CMD_PATH_LEN,
2536		.mode		= 0644,
2537		.proc_handler	= proc_dostring,
2538	},
2539#ifdef CONFIG_KEYS
2540	{
2541		.procname	= "keys",
2542		.mode		= 0555,
2543		.child		= key_sysctls,
2544	},
2545#endif
2546#ifdef CONFIG_PERF_EVENTS
2547	/*
2548	 * User-space scripts rely on the existence of this file
2549	 * as a feature check for perf_events being enabled.
2550	 *
2551	 * So it's an ABI, do not remove!
2552	 */
2553	{
2554		.procname	= "perf_event_paranoid",
2555		.data		= &sysctl_perf_event_paranoid,
2556		.maxlen		= sizeof(sysctl_perf_event_paranoid),
2557		.mode		= 0644,
2558		.proc_handler	= proc_dointvec,
2559	},
2560	{
2561		.procname	= "perf_event_mlock_kb",
2562		.data		= &sysctl_perf_event_mlock,
2563		.maxlen		= sizeof(sysctl_perf_event_mlock),
2564		.mode		= 0644,
2565		.proc_handler	= proc_dointvec,
2566	},
2567	{
2568		.procname	= "perf_event_max_sample_rate",
2569		.data		= &sysctl_perf_event_sample_rate,
2570		.maxlen		= sizeof(sysctl_perf_event_sample_rate),
2571		.mode		= 0644,
2572		.proc_handler	= perf_proc_update_handler,
2573		.extra1		= SYSCTL_ONE,
2574	},
2575	{
2576		.procname	= "perf_cpu_time_max_percent",
2577		.data		= &sysctl_perf_cpu_time_max_percent,
2578		.maxlen		= sizeof(sysctl_perf_cpu_time_max_percent),
2579		.mode		= 0644,
2580		.proc_handler	= perf_cpu_time_max_percent_handler,
2581		.extra1		= SYSCTL_ZERO,
2582		.extra2		= &one_hundred,
2583	},
2584	{
2585		.procname	= "perf_event_max_stack",
2586		.data		= &sysctl_perf_event_max_stack,
2587		.maxlen		= sizeof(sysctl_perf_event_max_stack),
2588		.mode		= 0644,
2589		.proc_handler	= perf_event_max_stack_handler,
2590		.extra1		= SYSCTL_ZERO,
2591		.extra2		= &six_hundred_forty_kb,
2592	},
2593	{
2594		.procname	= "perf_event_max_contexts_per_stack",
2595		.data		= &sysctl_perf_event_max_contexts_per_stack,
2596		.maxlen		= sizeof(sysctl_perf_event_max_contexts_per_stack),
2597		.mode		= 0644,
2598		.proc_handler	= perf_event_max_stack_handler,
2599		.extra1		= SYSCTL_ZERO,
2600		.extra2		= &one_thousand,
2601	},
2602#endif
2603	{
2604		.procname	= "panic_on_warn",
2605		.data		= &panic_on_warn,
2606		.maxlen		= sizeof(int),
2607		.mode		= 0644,
2608		.proc_handler	= proc_dointvec_minmax,
2609		.extra1		= SYSCTL_ZERO,
2610		.extra2		= SYSCTL_ONE,
2611	},
2612#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
2613	{
2614		.procname	= "timer_migration",
2615		.data		= &sysctl_timer_migration,
2616		.maxlen		= sizeof(unsigned int),
2617		.mode		= 0644,
2618		.proc_handler	= timer_migration_handler,
2619		.extra1		= SYSCTL_ZERO,
2620		.extra2		= SYSCTL_ONE,
2621	},
2622#endif
2623#ifdef CONFIG_BPF_SYSCALL
2624	{
2625		.procname	= "unprivileged_bpf_disabled",
2626		.data		= &sysctl_unprivileged_bpf_disabled,
2627		.maxlen		= sizeof(sysctl_unprivileged_bpf_disabled),
2628		.mode		= 0644,
2629		/* only handle a transition from default "0" to "1" */
2630		.proc_handler	= proc_dointvec_minmax,
2631		.extra1		= SYSCTL_ONE,
2632		.extra2		= SYSCTL_ONE,
2633	},
2634	{
2635		.procname	= "bpf_stats_enabled",
2636		.data		= &bpf_stats_enabled_key.key,
2637		.maxlen		= sizeof(bpf_stats_enabled_key),
2638		.mode		= 0644,
2639		.proc_handler	= bpf_stats_handler,
2640	},
2641#endif
2642#if defined(CONFIG_TREE_RCU)
2643	{
2644		.procname	= "panic_on_rcu_stall",
2645		.data		= &sysctl_panic_on_rcu_stall,
2646		.maxlen		= sizeof(sysctl_panic_on_rcu_stall),
2647		.mode		= 0644,
2648		.proc_handler	= proc_dointvec_minmax,
2649		.extra1		= SYSCTL_ZERO,
2650		.extra2		= SYSCTL_ONE,
2651	},
2652#endif
2653#ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
2654	{
2655		.procname	= "stack_erasing",
2656		.data		= NULL,
2657		.maxlen		= sizeof(int),
2658		.mode		= 0600,
2659		.proc_handler	= stack_erasing_sysctl,
2660		.extra1		= SYSCTL_ZERO,
2661		.extra2		= SYSCTL_ONE,
2662	},
2663#endif
2664	{ }
2665};
2666
2667static struct ctl_table vm_table[] = {
2668	{
2669		.procname	= "overcommit_memory",
2670		.data		= &sysctl_overcommit_memory,
2671		.maxlen		= sizeof(sysctl_overcommit_memory),
2672		.mode		= 0644,
2673		.proc_handler	= overcommit_policy_handler,
2674		.extra1		= SYSCTL_ZERO,
2675		.extra2		= &two,
2676	},
2677	{
2678		.procname	= "panic_on_oom",
2679		.data		= &sysctl_panic_on_oom,
2680		.maxlen		= sizeof(sysctl_panic_on_oom),
2681		.mode		= 0644,
2682		.proc_handler	= proc_dointvec_minmax,
2683		.extra1		= SYSCTL_ZERO,
2684		.extra2		= &two,
2685	},
2686	{
2687		.procname	= "oom_kill_allocating_task",
2688		.data		= &sysctl_oom_kill_allocating_task,
2689		.maxlen		= sizeof(sysctl_oom_kill_allocating_task),
2690		.mode		= 0644,
2691		.proc_handler	= proc_dointvec,
2692	},
2693	{
2694		.procname	= "oom_dump_tasks",
2695		.data		= &sysctl_oom_dump_tasks,
2696		.maxlen		= sizeof(sysctl_oom_dump_tasks),
2697		.mode		= 0644,
2698		.proc_handler	= proc_dointvec,
2699	},
2700	{
2701		.procname	= "overcommit_ratio",
2702		.data		= &sysctl_overcommit_ratio,
2703		.maxlen		= sizeof(sysctl_overcommit_ratio),
2704		.mode		= 0644,
2705		.proc_handler	= overcommit_ratio_handler,
2706	},
2707	{
2708		.procname	= "overcommit_kbytes",
2709		.data		= &sysctl_overcommit_kbytes,
2710		.maxlen		= sizeof(sysctl_overcommit_kbytes),
2711		.mode		= 0644,
2712		.proc_handler	= overcommit_kbytes_handler,
2713	},
2714	{
2715		.procname	= "page-cluster",
2716		.data		= &page_cluster,
2717		.maxlen		= sizeof(int),
2718		.mode		= 0644,
2719		.proc_handler	= proc_dointvec_minmax,
2720		.extra1		= SYSCTL_ZERO,
2721	},
2722	{
2723		.procname	= "dirty_background_ratio",
2724		.data		= &dirty_background_ratio,
2725		.maxlen		= sizeof(dirty_background_ratio),
2726		.mode		= 0644,
2727		.proc_handler	= dirty_background_ratio_handler,
2728		.extra1		= SYSCTL_ZERO,
2729		.extra2		= &one_hundred,
2730	},
2731	{
2732		.procname	= "dirty_background_bytes",
2733		.data		= &dirty_background_bytes,
2734		.maxlen		= sizeof(dirty_background_bytes),
2735		.mode		= 0644,
2736		.proc_handler	= dirty_background_bytes_handler,
2737		.extra1		= &one_ul,
2738	},
2739	{
2740		.procname	= "dirty_ratio",
2741		.data		= &vm_dirty_ratio,
2742		.maxlen		= sizeof(vm_dirty_ratio),
2743		.mode		= 0644,
2744		.proc_handler	= dirty_ratio_handler,
2745		.extra1		= SYSCTL_ZERO,
2746		.extra2		= &one_hundred,
2747	},
2748	{
2749		.procname	= "dirty_bytes",
2750		.data		= &vm_dirty_bytes,
2751		.maxlen		= sizeof(vm_dirty_bytes),
2752		.mode		= 0644,
2753		.proc_handler	= dirty_bytes_handler,
2754		.extra1		= &dirty_bytes_min,
2755	},
2756	{
2757		.procname	= "dirty_writeback_centisecs",
2758		.data		= &dirty_writeback_interval,
2759		.maxlen		= sizeof(dirty_writeback_interval),
2760		.mode		= 0644,
2761		.proc_handler	= dirty_writeback_centisecs_handler,
2762	},
2763	{
2764		.procname	= "dirty_expire_centisecs",
2765		.data		= &dirty_expire_interval,
2766		.maxlen		= sizeof(dirty_expire_interval),
2767		.mode		= 0644,
2768		.proc_handler	= proc_dointvec_minmax,
2769		.extra1		= SYSCTL_ZERO,
2770	},
2771	{
2772		.procname	= "dirtytime_expire_seconds",
2773		.data		= &dirtytime_expire_interval,
2774		.maxlen		= sizeof(dirtytime_expire_interval),
2775		.mode		= 0644,
2776		.proc_handler	= dirtytime_interval_handler,
2777		.extra1		= SYSCTL_ZERO,
2778	},
2779	{
2780		.procname	= "swappiness",
2781		.data		= &vm_swappiness,
2782		.maxlen		= sizeof(vm_swappiness),
2783		.mode		= 0644,
2784		.proc_handler	= proc_dointvec_minmax,
2785		.extra1		= SYSCTL_ZERO,
2786		.extra2		= &two_hundred,
2787	},
2788#ifdef CONFIG_HUGETLB_PAGE
2789	{
2790		.procname	= "nr_hugepages",
2791		.data		= NULL,
2792		.maxlen		= sizeof(unsigned long),
2793		.mode		= 0644,
2794		.proc_handler	= hugetlb_sysctl_handler,
2795	},
2796#ifdef CONFIG_NUMA
2797	{
2798		.procname       = "nr_hugepages_mempolicy",
2799		.data           = NULL,
2800		.maxlen         = sizeof(unsigned long),
2801		.mode           = 0644,
2802		.proc_handler   = &hugetlb_mempolicy_sysctl_handler,
2803	},
2804	{
2805		.procname		= "numa_stat",
2806		.data			= &sysctl_vm_numa_stat,
2807		.maxlen			= sizeof(int),
2808		.mode			= 0644,
2809		.proc_handler	= sysctl_vm_numa_stat_handler,
2810		.extra1			= SYSCTL_ZERO,
2811		.extra2			= SYSCTL_ONE,
2812	},
2813#endif
2814	 {
2815		.procname	= "hugetlb_shm_group",
2816		.data		= &sysctl_hugetlb_shm_group,
2817		.maxlen		= sizeof(gid_t),
2818		.mode		= 0644,
2819		.proc_handler	= proc_dointvec,
2820	 },
2821	{
2822		.procname	= "nr_overcommit_hugepages",
2823		.data		= NULL,
2824		.maxlen		= sizeof(unsigned long),
2825		.mode		= 0644,
2826		.proc_handler	= hugetlb_overcommit_handler,
2827	},
2828#endif
2829	{
2830		.procname	= "lowmem_reserve_ratio",
2831		.data		= &sysctl_lowmem_reserve_ratio,
2832		.maxlen		= sizeof(sysctl_lowmem_reserve_ratio),
2833		.mode		= 0644,
2834		.proc_handler	= lowmem_reserve_ratio_sysctl_handler,
2835	},
2836	{
2837		.procname	= "drop_caches",
2838		.data		= &sysctl_drop_caches,
2839		.maxlen		= sizeof(int),
2840		.mode		= 0200,
2841		.proc_handler	= drop_caches_sysctl_handler,
2842		.extra1		= SYSCTL_ONE,
2843		.extra2		= &four,
2844	},
2845#ifdef CONFIG_COMPACTION
2846	{
2847		.procname	= "compact_memory",
2848		.data		= &sysctl_compact_memory,
2849		.maxlen		= sizeof(int),
2850		.mode		= 0200,
2851		.proc_handler	= sysctl_compaction_handler,
2852	},
2853	{
2854		.procname	= "compaction_proactiveness",
2855		.data		= &sysctl_compaction_proactiveness,
2856		.maxlen		= sizeof(sysctl_compaction_proactiveness),
2857		.mode		= 0644,
2858		.proc_handler	= proc_dointvec_minmax,
2859		.extra1		= SYSCTL_ZERO,
2860		.extra2		= &one_hundred,
2861	},
2862	{
2863		.procname	= "extfrag_threshold",
2864		.data		= &sysctl_extfrag_threshold,
2865		.maxlen		= sizeof(int),
2866		.mode		= 0644,
2867		.proc_handler	= proc_dointvec_minmax,
2868		.extra1		= &min_extfrag_threshold,
2869		.extra2		= &max_extfrag_threshold,
2870	},
2871	{
2872		.procname	= "compact_unevictable_allowed",
2873		.data		= &sysctl_compact_unevictable_allowed,
2874		.maxlen		= sizeof(int),
2875		.mode		= 0644,
2876		.proc_handler	= proc_dointvec_minmax_warn_RT_change,
2877		.extra1		= SYSCTL_ZERO,
2878		.extra2		= SYSCTL_ONE,
2879	},
2880
2881#endif /* CONFIG_COMPACTION */
2882	{
2883		.procname	= "min_free_kbytes",
2884		.data		= &min_free_kbytes,
2885		.maxlen		= sizeof(min_free_kbytes),
2886		.mode		= 0644,
2887		.proc_handler	= min_free_kbytes_sysctl_handler,
2888		.extra1		= SYSCTL_ZERO,
2889	},
2890	{
2891		.procname	= "watermark_boost_factor",
2892		.data		= &watermark_boost_factor,
2893		.maxlen		= sizeof(watermark_boost_factor),
2894		.mode		= 0644,
2895		.proc_handler	= proc_dointvec_minmax,
2896		.extra1		= SYSCTL_ZERO,
2897	},
2898	{
2899		.procname	= "watermark_scale_factor",
2900		.data		= &watermark_scale_factor,
2901		.maxlen		= sizeof(watermark_scale_factor),
2902		.mode		= 0644,
2903		.proc_handler	= watermark_scale_factor_sysctl_handler,
2904		.extra1		= SYSCTL_ONE,
2905		.extra2		= &one_thousand,
2906	},
2907	{
2908		.procname	= "percpu_pagelist_fraction",
2909		.data		= &percpu_pagelist_fraction,
2910		.maxlen		= sizeof(percpu_pagelist_fraction),
2911		.mode		= 0644,
2912		.proc_handler	= percpu_pagelist_fraction_sysctl_handler,
2913		.extra1		= SYSCTL_ZERO,
2914	},
2915	{
2916		.procname	= "page_lock_unfairness",
2917		.data		= &sysctl_page_lock_unfairness,
2918		.maxlen		= sizeof(sysctl_page_lock_unfairness),
2919		.mode		= 0644,
2920		.proc_handler	= proc_dointvec_minmax,
2921		.extra1		= SYSCTL_ZERO,
2922	},
2923#ifdef CONFIG_MMU
2924	{
2925		.procname	= "max_map_count",
2926		.data		= &sysctl_max_map_count,
2927		.maxlen		= sizeof(sysctl_max_map_count),
2928		.mode		= 0644,
2929		.proc_handler	= proc_dointvec_minmax,
2930		.extra1		= SYSCTL_ZERO,
2931	},
2932#else
2933	{
2934		.procname	= "nr_trim_pages",
2935		.data		= &sysctl_nr_trim_pages,
2936		.maxlen		= sizeof(sysctl_nr_trim_pages),
2937		.mode		= 0644,
2938		.proc_handler	= proc_dointvec_minmax,
2939		.extra1		= SYSCTL_ZERO,
2940	},
2941#endif
2942	{
2943		.procname	= "laptop_mode",
2944		.data		= &laptop_mode,
2945		.maxlen		= sizeof(laptop_mode),
2946		.mode		= 0644,
2947		.proc_handler	= proc_dointvec_jiffies,
2948	},
2949	{
2950		.procname	= "block_dump",
2951		.data		= &block_dump,
2952		.maxlen		= sizeof(block_dump),
2953		.mode		= 0644,
2954		.proc_handler	= proc_dointvec,
2955		.extra1		= SYSCTL_ZERO,
2956	},
2957	{
2958		.procname	= "vfs_cache_pressure",
2959		.data		= &sysctl_vfs_cache_pressure,
2960		.maxlen		= sizeof(sysctl_vfs_cache_pressure),
2961		.mode		= 0644,
2962		.proc_handler	= proc_dointvec,
2963		.extra1		= SYSCTL_ZERO,
2964	},
2965#if defined(HAVE_ARCH_PICK_MMAP_LAYOUT) || \
2966    defined(CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT)
2967	{
2968		.procname	= "legacy_va_layout",
2969		.data		= &sysctl_legacy_va_layout,
2970		.maxlen		= sizeof(sysctl_legacy_va_layout),
2971		.mode		= 0644,
2972		.proc_handler	= proc_dointvec,
2973		.extra1		= SYSCTL_ZERO,
2974	},
2975#endif
2976#ifdef CONFIG_NUMA
2977	{
2978		.procname	= "zone_reclaim_mode",
2979		.data		= &node_reclaim_mode,
2980		.maxlen		= sizeof(node_reclaim_mode),
2981		.mode		= 0644,
2982		.proc_handler	= proc_dointvec,
2983		.extra1		= SYSCTL_ZERO,
2984	},
2985	{
2986		.procname	= "min_unmapped_ratio",
2987		.data		= &sysctl_min_unmapped_ratio,
2988		.maxlen		= sizeof(sysctl_min_unmapped_ratio),
2989		.mode		= 0644,
2990		.proc_handler	= sysctl_min_unmapped_ratio_sysctl_handler,
2991		.extra1		= SYSCTL_ZERO,
2992		.extra2		= &one_hundred,
2993	},
2994	{
2995		.procname	= "min_slab_ratio",
2996		.data		= &sysctl_min_slab_ratio,
2997		.maxlen		= sizeof(sysctl_min_slab_ratio),
2998		.mode		= 0644,
2999		.proc_handler	= sysctl_min_slab_ratio_sysctl_handler,
3000		.extra1		= SYSCTL_ZERO,
3001		.extra2		= &one_hundred,
3002	},
3003#endif
3004#ifdef CONFIG_SMP
3005	{
3006		.procname	= "stat_interval",
3007		.data		= &sysctl_stat_interval,
3008		.maxlen		= sizeof(sysctl_stat_interval),
3009		.mode		= 0644,
3010		.proc_handler	= proc_dointvec_jiffies,
3011	},
3012	{
3013		.procname	= "stat_refresh",
3014		.data		= NULL,
3015		.maxlen		= 0,
3016		.mode		= 0600,
3017		.proc_handler	= vmstat_refresh,
3018	},
3019#endif
3020#ifdef CONFIG_MMU
3021	{
3022		.procname	= "mmap_min_addr",
3023		.data		= &dac_mmap_min_addr,
3024		.maxlen		= sizeof(unsigned long),
3025		.mode		= 0644,
3026		.proc_handler	= mmap_min_addr_handler,
3027	},
3028#endif
3029#ifdef CONFIG_NUMA
3030	{
3031		.procname	= "numa_zonelist_order",
3032		.data		= &numa_zonelist_order,
3033		.maxlen		= NUMA_ZONELIST_ORDER_LEN,
3034		.mode		= 0644,
3035		.proc_handler	= numa_zonelist_order_handler,
3036	},
3037#endif
3038#if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
3039   (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
3040	{
3041		.procname	= "vdso_enabled",
3042#ifdef CONFIG_X86_32
3043		.data		= &vdso32_enabled,
3044		.maxlen		= sizeof(vdso32_enabled),
3045#else
3046		.data		= &vdso_enabled,
3047		.maxlen		= sizeof(vdso_enabled),
3048#endif
3049		.mode		= 0644,
3050		.proc_handler	= proc_dointvec,
3051		.extra1		= SYSCTL_ZERO,
3052	},
3053#endif
3054#ifdef CONFIG_HIGHMEM
3055	{
3056		.procname	= "highmem_is_dirtyable",
3057		.data		= &vm_highmem_is_dirtyable,
3058		.maxlen		= sizeof(vm_highmem_is_dirtyable),
3059		.mode		= 0644,
3060		.proc_handler	= proc_dointvec_minmax,
3061		.extra1		= SYSCTL_ZERO,
3062		.extra2		= SYSCTL_ONE,
3063	},
3064#endif
3065#ifdef CONFIG_MEMORY_FAILURE
3066	{
3067		.procname	= "memory_failure_early_kill",
3068		.data		= &sysctl_memory_failure_early_kill,
3069		.maxlen		= sizeof(sysctl_memory_failure_early_kill),
3070		.mode		= 0644,
3071		.proc_handler	= proc_dointvec_minmax,
3072		.extra1		= SYSCTL_ZERO,
3073		.extra2		= SYSCTL_ONE,
3074	},
3075	{
3076		.procname	= "memory_failure_recovery",
3077		.data		= &sysctl_memory_failure_recovery,
3078		.maxlen		= sizeof(sysctl_memory_failure_recovery),
3079		.mode		= 0644,
3080		.proc_handler	= proc_dointvec_minmax,
3081		.extra1		= SYSCTL_ZERO,
3082		.extra2		= SYSCTL_ONE,
3083	},
3084#endif
3085	{
3086		.procname	= "user_reserve_kbytes",
3087		.data		= &sysctl_user_reserve_kbytes,
3088		.maxlen		= sizeof(sysctl_user_reserve_kbytes),
3089		.mode		= 0644,
3090		.proc_handler	= proc_doulongvec_minmax,
3091	},
3092	{
3093		.procname	= "admin_reserve_kbytes",
3094		.data		= &sysctl_admin_reserve_kbytes,
3095		.maxlen		= sizeof(sysctl_admin_reserve_kbytes),
3096		.mode		= 0644,
3097		.proc_handler	= proc_doulongvec_minmax,
3098	},
3099#ifdef CONFIG_HAVE_ARCH_MMAP_RND_BITS
3100	{
3101		.procname	= "mmap_rnd_bits",
3102		.data		= &mmap_rnd_bits,
3103		.maxlen		= sizeof(mmap_rnd_bits),
3104		.mode		= 0600,
3105		.proc_handler	= proc_dointvec_minmax,
3106		.extra1		= (void *)&mmap_rnd_bits_min,
3107		.extra2		= (void *)&mmap_rnd_bits_max,
3108	},
3109#endif
3110#ifdef CONFIG_HAVE_ARCH_MMAP_RND_COMPAT_BITS
3111	{
3112		.procname	= "mmap_rnd_compat_bits",
3113		.data		= &mmap_rnd_compat_bits,
3114		.maxlen		= sizeof(mmap_rnd_compat_bits),
3115		.mode		= 0600,
3116		.proc_handler	= proc_dointvec_minmax,
3117		.extra1		= (void *)&mmap_rnd_compat_bits_min,
3118		.extra2		= (void *)&mmap_rnd_compat_bits_max,
3119	},
3120#endif
3121#ifdef CONFIG_USERFAULTFD
3122	{
3123		.procname	= "unprivileged_userfaultfd",
3124		.data		= &sysctl_unprivileged_userfaultfd,
3125		.maxlen		= sizeof(sysctl_unprivileged_userfaultfd),
3126		.mode		= 0644,
3127		.proc_handler	= proc_dointvec_minmax,
3128		.extra1		= SYSCTL_ZERO,
3129		.extra2		= SYSCTL_ONE,
3130	},
3131#endif
3132	{ }
3133};
3134
3135static struct ctl_table fs_table[] = {
3136	{
3137		.procname	= "inode-nr",
3138		.data		= &inodes_stat,
3139		.maxlen		= 2*sizeof(long),
3140		.mode		= 0444,
3141		.proc_handler	= proc_nr_inodes,
3142	},
3143	{
3144		.procname	= "inode-state",
3145		.data		= &inodes_stat,
3146		.maxlen		= 7*sizeof(long),
3147		.mode		= 0444,
3148		.proc_handler	= proc_nr_inodes,
3149	},
3150	{
3151		.procname	= "file-nr",
3152		.data		= &files_stat,
3153		.maxlen		= sizeof(files_stat),
3154		.mode		= 0444,
3155		.proc_handler	= proc_nr_files,
3156	},
3157	{
3158		.procname	= "file-max",
3159		.data		= &files_stat.max_files,
3160		.maxlen		= sizeof(files_stat.max_files),
3161		.mode		= 0644,
3162		.proc_handler	= proc_doulongvec_minmax,
3163		.extra1		= &zero_ul,
3164		.extra2		= &long_max,
3165	},
3166	{
3167		.procname	= "nr_open",
3168		.data		= &sysctl_nr_open,
3169		.maxlen		= sizeof(unsigned int),
3170		.mode		= 0644,
3171		.proc_handler	= proc_dointvec_minmax,
3172		.extra1		= &sysctl_nr_open_min,
3173		.extra2		= &sysctl_nr_open_max,
3174	},
3175	{
3176		.procname	= "dentry-state",
3177		.data		= &dentry_stat,
3178		.maxlen		= 6*sizeof(long),
3179		.mode		= 0444,
3180		.proc_handler	= proc_nr_dentry,
3181	},
3182	{
3183		.procname	= "overflowuid",
3184		.data		= &fs_overflowuid,
3185		.maxlen		= sizeof(int),
3186		.mode		= 0644,
3187		.proc_handler	= proc_dointvec_minmax,
3188		.extra1		= &minolduid,
3189		.extra2		= &maxolduid,
3190	},
3191	{
3192		.procname	= "overflowgid",
3193		.data		= &fs_overflowgid,
3194		.maxlen		= sizeof(int),
3195		.mode		= 0644,
3196		.proc_handler	= proc_dointvec_minmax,
3197		.extra1		= &minolduid,
3198		.extra2		= &maxolduid,
3199	},
3200#ifdef CONFIG_FILE_LOCKING
3201	{
3202		.procname	= "leases-enable",
3203		.data		= &leases_enable,
3204		.maxlen		= sizeof(int),
3205		.mode		= 0644,
3206		.proc_handler	= proc_dointvec,
3207	},
3208#endif
3209#ifdef CONFIG_DNOTIFY
3210	{
3211		.procname	= "dir-notify-enable",
3212		.data		= &dir_notify_enable,
3213		.maxlen		= sizeof(int),
3214		.mode		= 0644,
3215		.proc_handler	= proc_dointvec,
3216	},
3217#endif
3218#ifdef CONFIG_MMU
3219#ifdef CONFIG_FILE_LOCKING
3220	{
3221		.procname	= "lease-break-time",
3222		.data		= &lease_break_time,
3223		.maxlen		= sizeof(int),
3224		.mode		= 0644,
3225		.proc_handler	= proc_dointvec,
3226	},
3227#endif
3228#ifdef CONFIG_AIO
3229	{
3230		.procname	= "aio-nr",
3231		.data		= &aio_nr,
3232		.maxlen		= sizeof(aio_nr),
3233		.mode		= 0444,
3234		.proc_handler	= proc_doulongvec_minmax,
3235	},
3236	{
3237		.procname	= "aio-max-nr",
3238		.data		= &aio_max_nr,
3239		.maxlen		= sizeof(aio_max_nr),
3240		.mode		= 0644,
3241		.proc_handler	= proc_doulongvec_minmax,
3242	},
3243#endif /* CONFIG_AIO */
3244#ifdef CONFIG_INOTIFY_USER
3245	{
3246		.procname	= "inotify",
3247		.mode		= 0555,
3248		.child		= inotify_table,
3249	},
3250#endif	
3251#ifdef CONFIG_EPOLL
3252	{
3253		.procname	= "epoll",
3254		.mode		= 0555,
3255		.child		= epoll_table,
3256	},
3257#endif
3258#endif
3259	{
3260		.procname	= "protected_symlinks",
3261		.data		= &sysctl_protected_symlinks,
3262		.maxlen		= sizeof(int),
3263		.mode		= 0600,
3264		.proc_handler	= proc_dointvec_minmax,
3265		.extra1		= SYSCTL_ZERO,
3266		.extra2		= SYSCTL_ONE,
3267	},
3268	{
3269		.procname	= "protected_hardlinks",
3270		.data		= &sysctl_protected_hardlinks,
3271		.maxlen		= sizeof(int),
3272		.mode		= 0600,
3273		.proc_handler	= proc_dointvec_minmax,
3274		.extra1		= SYSCTL_ZERO,
3275		.extra2		= SYSCTL_ONE,
3276	},
3277	{
3278		.procname	= "protected_fifos",
3279		.data		= &sysctl_protected_fifos,
3280		.maxlen		= sizeof(int),
3281		.mode		= 0600,
3282		.proc_handler	= proc_dointvec_minmax,
3283		.extra1		= SYSCTL_ZERO,
3284		.extra2		= &two,
3285	},
3286	{
3287		.procname	= "protected_regular",
3288		.data		= &sysctl_protected_regular,
3289		.maxlen		= sizeof(int),
3290		.mode		= 0600,
3291		.proc_handler	= proc_dointvec_minmax,
3292		.extra1		= SYSCTL_ZERO,
3293		.extra2		= &two,
3294	},
3295	{
3296		.procname	= "suid_dumpable",
3297		.data		= &suid_dumpable,
3298		.maxlen		= sizeof(int),
3299		.mode		= 0644,
3300		.proc_handler	= proc_dointvec_minmax_coredump,
3301		.extra1		= SYSCTL_ZERO,
3302		.extra2		= &two,
3303	},
3304#if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
3305	{
3306		.procname	= "binfmt_misc",
3307		.mode		= 0555,
3308		.child		= sysctl_mount_point,
3309	},
3310#endif
3311	{
3312		.procname	= "pipe-max-size",
3313		.data		= &pipe_max_size,
3314		.maxlen		= sizeof(pipe_max_size),
3315		.mode		= 0644,
3316		.proc_handler	= proc_dopipe_max_size,
3317	},
3318	{
3319		.procname	= "pipe-user-pages-hard",
3320		.data		= &pipe_user_pages_hard,
3321		.maxlen		= sizeof(pipe_user_pages_hard),
3322		.mode		= 0644,
3323		.proc_handler	= proc_doulongvec_minmax,
3324	},
3325	{
3326		.procname	= "pipe-user-pages-soft",
3327		.data		= &pipe_user_pages_soft,
3328		.maxlen		= sizeof(pipe_user_pages_soft),
3329		.mode		= 0644,
3330		.proc_handler	= proc_doulongvec_minmax,
3331	},
3332	{
3333		.procname	= "mount-max",
3334		.data		= &sysctl_mount_max,
3335		.maxlen		= sizeof(unsigned int),
3336		.mode		= 0644,
3337		.proc_handler	= proc_dointvec_minmax,
3338		.extra1		= SYSCTL_ONE,
3339	},
3340	{ }
3341};
3342
3343static struct ctl_table debug_table[] = {
3344#ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
3345	{
3346		.procname	= "exception-trace",
3347		.data		= &show_unhandled_signals,
3348		.maxlen		= sizeof(int),
3349		.mode		= 0644,
3350		.proc_handler	= proc_dointvec
3351	},
3352#endif
3353#if defined(CONFIG_OPTPROBES)
3354	{
3355		.procname	= "kprobes-optimization",
3356		.data		= &sysctl_kprobes_optimization,
3357		.maxlen		= sizeof(int),
3358		.mode		= 0644,
3359		.proc_handler	= proc_kprobes_optimization_handler,
3360		.extra1		= SYSCTL_ZERO,
3361		.extra2		= SYSCTL_ONE,
3362	},
3363#endif
3364	{ }
3365};
3366
3367static struct ctl_table dev_table[] = {
3368	{ }
3369};
3370
3371static struct ctl_table sysctl_base_table[] = {
3372	{
3373		.procname	= "kernel",
3374		.mode		= 0555,
3375		.child		= kern_table,
3376	},
3377	{
3378		.procname	= "vm",
3379		.mode		= 0555,
3380		.child		= vm_table,
3381	},
3382	{
3383		.procname	= "fs",
3384		.mode		= 0555,
3385		.child		= fs_table,
3386	},
3387	{
3388		.procname	= "debug",
3389		.mode		= 0555,
3390		.child		= debug_table,
3391	},
3392	{
3393		.procname	= "dev",
3394		.mode		= 0555,
3395		.child		= dev_table,
3396	},
3397	{ }
3398};
3399
3400int __init sysctl_init(void)
3401{
3402	struct ctl_table_header *hdr;
3403
3404	hdr = register_sysctl_table(sysctl_base_table);
3405	kmemleak_not_leak(hdr);
3406	return 0;
3407}
 
3408#endif /* CONFIG_SYSCTL */
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3409/*
3410 * No sense putting this after each symbol definition, twice,
3411 * exception granted :-)
3412 */
3413EXPORT_SYMBOL(proc_dointvec);
3414EXPORT_SYMBOL(proc_douintvec);
3415EXPORT_SYMBOL(proc_dointvec_jiffies);
3416EXPORT_SYMBOL(proc_dointvec_minmax);
3417EXPORT_SYMBOL_GPL(proc_douintvec_minmax);
3418EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
3419EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
3420EXPORT_SYMBOL(proc_dostring);
3421EXPORT_SYMBOL(proc_doulongvec_minmax);
3422EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
3423EXPORT_SYMBOL(proc_do_large_bitmap);