1/*
   2 * sysctl.c: General linux system control interface
   3 *
   4 * Begun 24 March 1995, Stephen Tweedie
   5 * Added /proc support, Dec 1995
   6 * Added bdflush entry and intvec min/max checking, 2/23/96, Tom Dyas.
   7 * Added hooks for /proc/sys/net (minor, minor patch), 96/4/1, Mike Shaver.
   8 * Added kernel/java-{interpreter,appletviewer}, 96/5/10, Mike Shaver.
   9 * Dynamic registration fixes, Stephen Tweedie.
  10 * Added kswapd-interval, ctrl-alt-del, printk stuff, 1/8/97, Chris Horn.
  11 * Made sysctl support optional via CONFIG_SYSCTL, 1/10/97, Chris
  12 *  Horn.
  13 * Added proc_doulongvec_ms_jiffies_minmax, 09/08/99, Carlos H. Bauer.
  14 * Added proc_doulongvec_minmax, 09/08/99, Carlos H. Bauer.
  15 * Changed linked lists to use list.h instead of lists.h, 02/24/00, Bill
  16 *  Wendling.
  17 * The list_for_each() macro wasn't appropriate for the sysctl loop.
  18 *  Removed it and replaced it with older style, 03/23/00, Bill Wendling
  19 */
  20
  21#include <linux/module.h>
  22#include <linux/mm.h>
  23#include <linux/swap.h>
  24#include <linux/slab.h>
  25#include <linux/sysctl.h>
  26#include <linux/bitmap.h>
  27#include <linux/signal.h>
  28#include <linux/printk.h>
  29#include <linux/proc_fs.h>
  30#include <linux/security.h>
  31#include <linux/ctype.h>
  32#include <linux/kmemcheck.h>
  33#include <linux/kmemleak.h>
  34#include <linux/fs.h>
  35#include <linux/init.h>
  36#include <linux/kernel.h>
  37#include <linux/kobject.h>
  38#include <linux/net.h>
  39#include <linux/sysrq.h>
  40#include <linux/highuid.h>
  41#include <linux/writeback.h>
  42#include <linux/ratelimit.h>
  43#include <linux/compaction.h>
  44#include <linux/hugetlb.h>
  45#include <linux/initrd.h>
  46#include <linux/key.h>
  47#include <linux/times.h>
  48#include <linux/limits.h>
  49#include <linux/dcache.h>
  50#include <linux/dnotify.h>
  51#include <linux/syscalls.h>
  52#include <linux/vmstat.h>
  53#include <linux/nfs_fs.h>
  54#include <linux/acpi.h>
  55#include <linux/reboot.h>
  56#include <linux/ftrace.h>
  57#include <linux/perf_event.h>
  58#include <linux/kprobes.h>
  59#include <linux/pipe_fs_i.h>
  60#include <linux/oom.h>
  61#include <linux/kmod.h>
  62#include <linux/capability.h>
  63#include <linux/binfmts.h>
  64#include <linux/sched/sysctl.h>
  65#include <linux/kexec.h>
  66
  67#include <asm/uaccess.h>
  68#include <asm/processor.h>
  69
  70#ifdef CONFIG_X86
  71#include <asm/nmi.h>
  72#include <asm/stacktrace.h>
  73#include <asm/io.h>
  74#endif
  75#ifdef CONFIG_SPARC
  76#include <asm/setup.h>
  77#endif
  78#ifdef CONFIG_BSD_PROCESS_ACCT
  79#include <linux/acct.h>
  80#endif
  81#ifdef CONFIG_RT_MUTEXES
  82#include <linux/rtmutex.h>
  83#endif
  84#if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
  85#include <linux/lockdep.h>
  86#endif
  87#ifdef CONFIG_CHR_DEV_SG
  88#include <scsi/sg.h>
  89#endif
  90
  91#ifdef CONFIG_LOCKUP_DETECTOR
  92#include <linux/nmi.h>
  93#endif
  94
  95
  96#if defined(CONFIG_SYSCTL)
  97
  98/* External variables not in a header file. */
 
 
  99extern int max_threads;
 100extern int suid_dumpable;
 101#ifdef CONFIG_COREDUMP
 102extern int core_uses_pid;
 
 103extern char core_pattern[];
 104extern unsigned int core_pipe_limit;
 105#endif
 106extern int pid_max;
 
 107extern int pid_max_min, pid_max_max;
 
 108extern int percpu_pagelist_fraction;
 109extern int compat_log;
 110extern int latencytop_enabled;
 111extern int sysctl_nr_open_min, sysctl_nr_open_max;
 112#ifndef CONFIG_MMU
 113extern int sysctl_nr_trim_pages;
 114#endif
 
 
 
 115
 116/* Constants used for minimum and  maximum */
 117#ifdef CONFIG_LOCKUP_DETECTOR
 118static int sixty = 60;
 
 119#endif
 120
 121static int __maybe_unused neg_one = -1;
 122
 123static int zero;
 124static int __maybe_unused one = 1;
 125static int __maybe_unused two = 2;
 126static int __maybe_unused four = 4;
 127static unsigned long one_ul = 1;
 128static int one_hundred = 100;
 129#ifdef CONFIG_PRINTK
 130static int ten_thousand = 10000;
 131#endif
 132
 133/* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
 134static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
 135
 136/* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
 137static int maxolduid = 65535;
 138static int minolduid;
 139static int min_percpu_pagelist_fract = 8;
 140
 141static int ngroups_max = NGROUPS_MAX;
 142static const int cap_last_cap = CAP_LAST_CAP;
 143
 144/*this is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs */
 145#ifdef CONFIG_DETECT_HUNG_TASK
 146static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
 147#endif
 148
 149#ifdef CONFIG_INOTIFY_USER
 150#include <linux/inotify.h>
 151#endif
 152#ifdef CONFIG_SPARC
 153#endif
 154
 155#ifdef CONFIG_SPARC64
 156extern int sysctl_tsb_ratio;
 157#endif
 158
 159#ifdef __hppa__
 160extern int pwrsw_enabled;
 161#endif
 162
 163#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
 164extern int unaligned_enabled;
 165#endif
 166
 167#ifdef CONFIG_IA64
 168extern int unaligned_dump_stack;
 169#endif
 170
 171#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
 172extern int no_unaligned_warning;
 
 173#endif
 174
 175#ifdef CONFIG_PROC_SYSCTL
 176static int proc_do_cad_pid(struct ctl_table *table, int write,
 177		  void __user *buffer, size_t *lenp, loff_t *ppos);
 178static int proc_taint(struct ctl_table *table, int write,
 179			       void __user *buffer, size_t *lenp, loff_t *ppos);
 180#endif
 181
 182#ifdef CONFIG_PRINTK
 183static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
 184				void __user *buffer, size_t *lenp, loff_t *ppos);
 185#endif
 186
 187static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
 188		void __user *buffer, size_t *lenp, loff_t *ppos);
 189#ifdef CONFIG_COREDUMP
 190static int proc_dostring_coredump(struct ctl_table *table, int write,
 191		void __user *buffer, size_t *lenp, loff_t *ppos);
 192#endif
 193
 194#ifdef CONFIG_MAGIC_SYSRQ
 195/* Note: sysrq code uses it's own private copy */
 196static int __sysrq_enabled = CONFIG_MAGIC_SYSRQ_DEFAULT_ENABLE;
 197
 198static int sysrq_sysctl_handler(ctl_table *table, int write,
 199				void __user *buffer, size_t *lenp,
 200				loff_t *ppos)
 201{
 202	int error;
 203
 204	error = proc_dointvec(table, write, buffer, lenp, ppos);
 205	if (error)
 206		return error;
 207
 208	if (write)
 209		sysrq_toggle_support(__sysrq_enabled);
 210
 211	return 0;
 212}
 213
 214#endif
 215
 216static struct ctl_table kern_table[];
 217static struct ctl_table vm_table[];
 218static struct ctl_table fs_table[];
 219static struct ctl_table debug_table[];
 220static struct ctl_table dev_table[];
 221extern struct ctl_table random_table[];
 222#ifdef CONFIG_EPOLL
 223extern struct ctl_table epoll_table[];
 224#endif
 225
 226#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
 227int sysctl_legacy_va_layout;
 228#endif
 229
 230/* The default sysctl tables: */
 231
 232static struct ctl_table sysctl_base_table[] = {
 233	{
 234		.procname	= "kernel",
 235		.mode		= 0555,
 236		.child		= kern_table,
 237	},
 238	{
 239		.procname	= "vm",
 240		.mode		= 0555,
 241		.child		= vm_table,
 242	},
 243	{
 244		.procname	= "fs",
 245		.mode		= 0555,
 246		.child		= fs_table,
 247	},
 248	{
 249		.procname	= "debug",
 250		.mode		= 0555,
 251		.child		= debug_table,
 252	},
 253	{
 254		.procname	= "dev",
 255		.mode		= 0555,
 256		.child		= dev_table,
 257	},
 258	{ }
 259};
 260
 261#ifdef CONFIG_SCHED_DEBUG
 262static int min_sched_granularity_ns = 100000;		/* 100 usecs */
 263static int max_sched_granularity_ns = NSEC_PER_SEC;	/* 1 second */
 264static int min_wakeup_granularity_ns;			/* 0 usecs */
 265static int max_wakeup_granularity_ns = NSEC_PER_SEC;	/* 1 second */
 266#ifdef CONFIG_SMP
 267static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
 268static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
 269#endif /* CONFIG_SMP */
 270#endif /* CONFIG_SCHED_DEBUG */
 271
 272#ifdef CONFIG_COMPACTION
 273static int min_extfrag_threshold;
 274static int max_extfrag_threshold = 1000;
 275#endif
 276
 277static struct ctl_table kern_table[] = {
 278	{
 279		.procname	= "sched_child_runs_first",
 280		.data		= &sysctl_sched_child_runs_first,
 281		.maxlen		= sizeof(unsigned int),
 282		.mode		= 0644,
 283		.proc_handler	= proc_dointvec,
 284	},
 285#ifdef CONFIG_SCHED_DEBUG
 286	{
 287		.procname	= "sched_min_granularity_ns",
 288		.data		= &sysctl_sched_min_granularity,
 289		.maxlen		= sizeof(unsigned int),
 290		.mode		= 0644,
 291		.proc_handler	= sched_proc_update_handler,
 292		.extra1		= &min_sched_granularity_ns,
 293		.extra2		= &max_sched_granularity_ns,
 294	},
 295	{
 296		.procname	= "sched_latency_ns",
 297		.data		= &sysctl_sched_latency,
 298		.maxlen		= sizeof(unsigned int),
 299		.mode		= 0644,
 300		.proc_handler	= sched_proc_update_handler,
 301		.extra1		= &min_sched_granularity_ns,
 302		.extra2		= &max_sched_granularity_ns,
 303	},
 304	{
 305		.procname	= "sched_wakeup_granularity_ns",
 306		.data		= &sysctl_sched_wakeup_granularity,
 307		.maxlen		= sizeof(unsigned int),
 308		.mode		= 0644,
 309		.proc_handler	= sched_proc_update_handler,
 310		.extra1		= &min_wakeup_granularity_ns,
 311		.extra2		= &max_wakeup_granularity_ns,
 312	},
 313#ifdef CONFIG_SMP
 314	{
 315		.procname	= "sched_tunable_scaling",
 316		.data		= &sysctl_sched_tunable_scaling,
 317		.maxlen		= sizeof(enum sched_tunable_scaling),
 318		.mode		= 0644,
 319		.proc_handler	= sched_proc_update_handler,
 320		.extra1		= &min_sched_tunable_scaling,
 321		.extra2		= &max_sched_tunable_scaling,
 322	},
 323	{
 324		.procname	= "sched_migration_cost_ns",
 325		.data		= &sysctl_sched_migration_cost,
 326		.maxlen		= sizeof(unsigned int),
 327		.mode		= 0644,
 328		.proc_handler	= proc_dointvec,
 329	},
 330	{
 331		.procname	= "sched_nr_migrate",
 332		.data		= &sysctl_sched_nr_migrate,
 333		.maxlen		= sizeof(unsigned int),
 334		.mode		= 0644,
 335		.proc_handler	= proc_dointvec,
 336	},
 337	{
 338		.procname	= "sched_time_avg_ms",
 339		.data		= &sysctl_sched_time_avg,
 340		.maxlen		= sizeof(unsigned int),
 341		.mode		= 0644,
 342		.proc_handler	= proc_dointvec,
 343	},
 344	{
 345		.procname	= "sched_shares_window_ns",
 346		.data		= &sysctl_sched_shares_window,
 347		.maxlen		= sizeof(unsigned int),
 348		.mode		= 0644,
 349		.proc_handler	= proc_dointvec,
 350	},
 351	{
 352		.procname	= "timer_migration",
 353		.data		= &sysctl_timer_migration,
 354		.maxlen		= sizeof(unsigned int),
 355		.mode		= 0644,
 356		.proc_handler	= proc_dointvec_minmax,
 357		.extra1		= &zero,
 358		.extra2		= &one,
 359	},
 360#endif /* CONFIG_SMP */
 361#ifdef CONFIG_NUMA_BALANCING
 362	{
 363		.procname	= "numa_balancing_scan_delay_ms",
 364		.data		= &sysctl_numa_balancing_scan_delay,
 365		.maxlen		= sizeof(unsigned int),
 366		.mode		= 0644,
 367		.proc_handler	= proc_dointvec,
 368	},
 369	{
 370		.procname	= "numa_balancing_scan_period_min_ms",
 371		.data		= &sysctl_numa_balancing_scan_period_min,
 372		.maxlen		= sizeof(unsigned int),
 373		.mode		= 0644,
 374		.proc_handler	= proc_dointvec,
 375	},
 376	{
 377		.procname	= "numa_balancing_scan_period_max_ms",
 378		.data		= &sysctl_numa_balancing_scan_period_max,
 379		.maxlen		= sizeof(unsigned int),
 380		.mode		= 0644,
 381		.proc_handler	= proc_dointvec,
 382	},
 383	{
 384		.procname	= "numa_balancing_scan_size_mb",
 385		.data		= &sysctl_numa_balancing_scan_size,
 386		.maxlen		= sizeof(unsigned int),
 387		.mode		= 0644,
 388		.proc_handler	= proc_dointvec,
 389	},
 390	{
 391		.procname	= "numa_balancing",
 392		.data		= NULL, /* filled in by handler */
 393		.maxlen		= sizeof(unsigned int),
 394		.mode		= 0644,
 395		.proc_handler	= sysctl_numa_balancing,
 396		.extra1		= &zero,
 397		.extra2		= &one,
 398	},
 399#endif /* CONFIG_NUMA_BALANCING */
 400#endif /* CONFIG_SCHED_DEBUG */
 401	{
 402		.procname	= "sched_rt_period_us",
 403		.data		= &sysctl_sched_rt_period,
 404		.maxlen		= sizeof(unsigned int),
 405		.mode		= 0644,
 406		.proc_handler	= sched_rt_handler,
 407	},
 408	{
 409		.procname	= "sched_rt_runtime_us",
 410		.data		= &sysctl_sched_rt_runtime,
 411		.maxlen		= sizeof(int),
 412		.mode		= 0644,
 413		.proc_handler	= sched_rt_handler,
 414	},
 415	{
 416		.procname	= "sched_rr_timeslice_ms",
 417		.data		= &sched_rr_timeslice,
 418		.maxlen		= sizeof(int),
 419		.mode		= 0644,
 420		.proc_handler	= sched_rr_handler,
 421	},
 422#ifdef CONFIG_SCHED_AUTOGROUP
 423	{
 424		.procname	= "sched_autogroup_enabled",
 425		.data		= &sysctl_sched_autogroup_enabled,
 426		.maxlen		= sizeof(unsigned int),
 427		.mode		= 0644,
 428		.proc_handler	= proc_dointvec_minmax,
 429		.extra1		= &zero,
 430		.extra2		= &one,
 431	},
 432#endif
 433#ifdef CONFIG_CFS_BANDWIDTH
 434	{
 435		.procname	= "sched_cfs_bandwidth_slice_us",
 436		.data		= &sysctl_sched_cfs_bandwidth_slice,
 437		.maxlen		= sizeof(unsigned int),
 438		.mode		= 0644,
 439		.proc_handler	= proc_dointvec_minmax,
 440		.extra1		= &one,
 441	},
 442#endif
 443#ifdef CONFIG_PROVE_LOCKING
 444	{
 445		.procname	= "prove_locking",
 446		.data		= &prove_locking,
 447		.maxlen		= sizeof(int),
 448		.mode		= 0644,
 449		.proc_handler	= proc_dointvec,
 450	},
 451#endif
 452#ifdef CONFIG_LOCK_STAT
 453	{
 454		.procname	= "lock_stat",
 455		.data		= &lock_stat,
 456		.maxlen		= sizeof(int),
 457		.mode		= 0644,
 458		.proc_handler	= proc_dointvec,
 459	},
 460#endif
 461	{
 462		.procname	= "panic",
 463		.data		= &panic_timeout,
 464		.maxlen		= sizeof(int),
 465		.mode		= 0644,
 466		.proc_handler	= proc_dointvec,
 467	},
 468#ifdef CONFIG_COREDUMP
 469	{
 470		.procname	= "core_uses_pid",
 471		.data		= &core_uses_pid,
 472		.maxlen		= sizeof(int),
 473		.mode		= 0644,
 474		.proc_handler	= proc_dointvec,
 475	},
 476	{
 477		.procname	= "core_pattern",
 478		.data		= core_pattern,
 479		.maxlen		= CORENAME_MAX_SIZE,
 480		.mode		= 0644,
 481		.proc_handler	= proc_dostring_coredump,
 482	},
 483	{
 484		.procname	= "core_pipe_limit",
 485		.data		= &core_pipe_limit,
 486		.maxlen		= sizeof(unsigned int),
 487		.mode		= 0644,
 488		.proc_handler	= proc_dointvec,
 489	},
 490#endif
 491#ifdef CONFIG_PROC_SYSCTL
 492	{
 493		.procname	= "tainted",
 494		.maxlen 	= sizeof(long),
 495		.mode		= 0644,
 496		.proc_handler	= proc_taint,
 497	},
 498#endif
 499#ifdef CONFIG_LATENCYTOP
 500	{
 501		.procname	= "latencytop",
 502		.data		= &latencytop_enabled,
 503		.maxlen		= sizeof(int),
 504		.mode		= 0644,
 505		.proc_handler	= proc_dointvec,
 506	},
 507#endif
 508#ifdef CONFIG_BLK_DEV_INITRD
 509	{
 510		.procname	= "real-root-dev",
 511		.data		= &real_root_dev,
 512		.maxlen		= sizeof(int),
 513		.mode		= 0644,
 514		.proc_handler	= proc_dointvec,
 515	},
 516#endif
 517	{
 518		.procname	= "print-fatal-signals",
 519		.data		= &print_fatal_signals,
 520		.maxlen		= sizeof(int),
 521		.mode		= 0644,
 522		.proc_handler	= proc_dointvec,
 523	},
 524#ifdef CONFIG_SPARC
 525	{
 526		.procname	= "reboot-cmd",
 527		.data		= reboot_command,
 528		.maxlen		= 256,
 529		.mode		= 0644,
 530		.proc_handler	= proc_dostring,
 531	},
 532	{
 533		.procname	= "stop-a",
 534		.data		= &stop_a_enabled,
 535		.maxlen		= sizeof (int),
 536		.mode		= 0644,
 537		.proc_handler	= proc_dointvec,
 538	},
 539	{
 540		.procname	= "scons-poweroff",
 541		.data		= &scons_pwroff,
 542		.maxlen		= sizeof (int),
 543		.mode		= 0644,
 544		.proc_handler	= proc_dointvec,
 545	},
 546#endif
 547#ifdef CONFIG_SPARC64
 548	{
 549		.procname	= "tsb-ratio",
 550		.data		= &sysctl_tsb_ratio,
 551		.maxlen		= sizeof (int),
 552		.mode		= 0644,
 553		.proc_handler	= proc_dointvec,
 554	},
 555#endif
 556#ifdef __hppa__
 557	{
 558		.procname	= "soft-power",
 559		.data		= &pwrsw_enabled,
 560		.maxlen		= sizeof (int),
 561	 	.mode		= 0644,
 562		.proc_handler	= proc_dointvec,
 563	},
 564#endif
 565#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
 566	{
 567		.procname	= "unaligned-trap",
 568		.data		= &unaligned_enabled,
 569		.maxlen		= sizeof (int),
 570		.mode		= 0644,
 571		.proc_handler	= proc_dointvec,
 572	},
 573#endif
 574	{
 575		.procname	= "ctrl-alt-del",
 576		.data		= &C_A_D,
 577		.maxlen		= sizeof(int),
 578		.mode		= 0644,
 579		.proc_handler	= proc_dointvec,
 580	},
 581#ifdef CONFIG_FUNCTION_TRACER
 582	{
 583		.procname	= "ftrace_enabled",
 584		.data		= &ftrace_enabled,
 585		.maxlen		= sizeof(int),
 586		.mode		= 0644,
 587		.proc_handler	= ftrace_enable_sysctl,
 588	},
 589#endif
 590#ifdef CONFIG_STACK_TRACER
 591	{
 592		.procname	= "stack_tracer_enabled",
 593		.data		= &stack_tracer_enabled,
 594		.maxlen		= sizeof(int),
 595		.mode		= 0644,
 596		.proc_handler	= stack_trace_sysctl,
 597	},
 598#endif
 599#ifdef CONFIG_TRACING
 600	{
 601		.procname	= "ftrace_dump_on_oops",
 602		.data		= &ftrace_dump_on_oops,
 603		.maxlen		= sizeof(int),
 604		.mode		= 0644,
 605		.proc_handler	= proc_dointvec,
 606	},
 607	{
 608		.procname	= "traceoff_on_warning",
 609		.data		= &__disable_trace_on_warning,
 610		.maxlen		= sizeof(__disable_trace_on_warning),
 611		.mode		= 0644,
 612		.proc_handler	= proc_dointvec,
 613	},
 614#endif
 615#ifdef CONFIG_KEXEC
 616	{
 617		.procname	= "kexec_load_disabled",
 618		.data		= &kexec_load_disabled,
 619		.maxlen		= sizeof(int),
 620		.mode		= 0644,
 621		/* only handle a transition from default "0" to "1" */
 622		.proc_handler	= proc_dointvec_minmax,
 623		.extra1		= &one,
 624		.extra2		= &one,
 625	},
 626#endif
 627#ifdef CONFIG_MODULES
 628	{
 629		.procname	= "modprobe",
 630		.data		= &modprobe_path,
 631		.maxlen		= KMOD_PATH_LEN,
 632		.mode		= 0644,
 633		.proc_handler	= proc_dostring,
 634	},
 635	{
 636		.procname	= "modules_disabled",
 637		.data		= &modules_disabled,
 638		.maxlen		= sizeof(int),
 639		.mode		= 0644,
 640		/* only handle a transition from default "0" to "1" */
 641		.proc_handler	= proc_dointvec_minmax,
 642		.extra1		= &one,
 643		.extra2		= &one,
 644	},
 645#endif
 646
 647	{
 648		.procname	= "hotplug",
 649		.data		= &uevent_helper,
 650		.maxlen		= UEVENT_HELPER_PATH_LEN,
 651		.mode		= 0644,
 652		.proc_handler	= proc_dostring,
 653	},
 654
 655#ifdef CONFIG_CHR_DEV_SG
 656	{
 657		.procname	= "sg-big-buff",
 658		.data		= &sg_big_buff,
 659		.maxlen		= sizeof (int),
 660		.mode		= 0444,
 661		.proc_handler	= proc_dointvec,
 662	},
 663#endif
 664#ifdef CONFIG_BSD_PROCESS_ACCT
 665	{
 666		.procname	= "acct",
 667		.data		= &acct_parm,
 668		.maxlen		= 3*sizeof(int),
 669		.mode		= 0644,
 670		.proc_handler	= proc_dointvec,
 671	},
 672#endif
 673#ifdef CONFIG_MAGIC_SYSRQ
 674	{
 675		.procname	= "sysrq",
 676		.data		= &__sysrq_enabled,
 677		.maxlen		= sizeof (int),
 678		.mode		= 0644,
 679		.proc_handler	= sysrq_sysctl_handler,
 680	},
 681#endif
 682#ifdef CONFIG_PROC_SYSCTL
 683	{
 684		.procname	= "cad_pid",
 685		.data		= NULL,
 686		.maxlen		= sizeof (int),
 687		.mode		= 0600,
 688		.proc_handler	= proc_do_cad_pid,
 689	},
 690#endif
 691	{
 692		.procname	= "threads-max",
 693		.data		= &max_threads,
 694		.maxlen		= sizeof(int),
 695		.mode		= 0644,
 696		.proc_handler	= proc_dointvec,
 697	},
 698	{
 699		.procname	= "random",
 700		.mode		= 0555,
 701		.child		= random_table,
 702	},
 703	{
 704		.procname	= "usermodehelper",
 705		.mode		= 0555,
 706		.child		= usermodehelper_table,
 707	},
 708	{
 709		.procname	= "overflowuid",
 710		.data		= &overflowuid,
 711		.maxlen		= sizeof(int),
 712		.mode		= 0644,
 713		.proc_handler	= proc_dointvec_minmax,
 714		.extra1		= &minolduid,
 715		.extra2		= &maxolduid,
 716	},
 717	{
 718		.procname	= "overflowgid",
 719		.data		= &overflowgid,
 720		.maxlen		= sizeof(int),
 721		.mode		= 0644,
 722		.proc_handler	= proc_dointvec_minmax,
 723		.extra1		= &minolduid,
 724		.extra2		= &maxolduid,
 725	},
 726#ifdef CONFIG_S390
 727#ifdef CONFIG_MATHEMU
 728	{
 729		.procname	= "ieee_emulation_warnings",
 730		.data		= &sysctl_ieee_emulation_warnings,
 731		.maxlen		= sizeof(int),
 732		.mode		= 0644,
 733		.proc_handler	= proc_dointvec,
 734	},
 735#endif
 736	{
 737		.procname	= "userprocess_debug",
 738		.data		= &show_unhandled_signals,
 739		.maxlen		= sizeof(int),
 740		.mode		= 0644,
 741		.proc_handler	= proc_dointvec,
 742	},
 743#endif
 744	{
 745		.procname	= "pid_max",
 746		.data		= &pid_max,
 747		.maxlen		= sizeof (int),
 748		.mode		= 0644,
 749		.proc_handler	= proc_dointvec_minmax,
 750		.extra1		= &pid_max_min,
 751		.extra2		= &pid_max_max,
 752	},
 753	{
 754		.procname	= "panic_on_oops",
 755		.data		= &panic_on_oops,
 756		.maxlen		= sizeof(int),
 757		.mode		= 0644,
 758		.proc_handler	= proc_dointvec,
 759	},
 760#if defined CONFIG_PRINTK
 761	{
 762		.procname	= "printk",
 763		.data		= &console_loglevel,
 764		.maxlen		= 4*sizeof(int),
 765		.mode		= 0644,
 766		.proc_handler	= proc_dointvec,
 767	},
 768	{
 769		.procname	= "printk_ratelimit",
 770		.data		= &printk_ratelimit_state.interval,
 771		.maxlen		= sizeof(int),
 772		.mode		= 0644,
 773		.proc_handler	= proc_dointvec_jiffies,
 774	},
 775	{
 776		.procname	= "printk_ratelimit_burst",
 777		.data		= &printk_ratelimit_state.burst,
 778		.maxlen		= sizeof(int),
 779		.mode		= 0644,
 780		.proc_handler	= proc_dointvec,
 781	},
 782	{
 783		.procname	= "printk_delay",
 784		.data		= &printk_delay_msec,
 785		.maxlen		= sizeof(int),
 786		.mode		= 0644,
 787		.proc_handler	= proc_dointvec_minmax,
 788		.extra1		= &zero,
 789		.extra2		= &ten_thousand,
 790	},
 791	{
 792		.procname	= "dmesg_restrict",
 793		.data		= &dmesg_restrict,
 794		.maxlen		= sizeof(int),
 795		.mode		= 0644,
 796		.proc_handler	= proc_dointvec_minmax_sysadmin,
 797		.extra1		= &zero,
 798		.extra2		= &one,
 799	},
 800	{
 801		.procname	= "kptr_restrict",
 802		.data		= &kptr_restrict,
 803		.maxlen		= sizeof(int),
 804		.mode		= 0644,
 805		.proc_handler	= proc_dointvec_minmax_sysadmin,
 806		.extra1		= &zero,
 807		.extra2		= &two,
 808	},
 809#endif
 810	{
 811		.procname	= "ngroups_max",
 812		.data		= &ngroups_max,
 813		.maxlen		= sizeof (int),
 814		.mode		= 0444,
 815		.proc_handler	= proc_dointvec,
 816	},
 817	{
 818		.procname	= "cap_last_cap",
 819		.data		= (void *)&cap_last_cap,
 820		.maxlen		= sizeof(int),
 821		.mode		= 0444,
 822		.proc_handler	= proc_dointvec,
 823	},
 824#if defined(CONFIG_LOCKUP_DETECTOR)
 825	{
 826		.procname       = "watchdog",
 827		.data           = &watchdog_user_enabled,
 828		.maxlen         = sizeof (int),
 829		.mode           = 0644,
 830		.proc_handler   = proc_dowatchdog,
 831		.extra1		= &zero,
 832		.extra2		= &one,
 833	},
 834	{
 835		.procname	= "watchdog_thresh",
 836		.data		= &watchdog_thresh,
 837		.maxlen		= sizeof(int),
 838		.mode		= 0644,
 839		.proc_handler	= proc_dowatchdog,
 840		.extra1		= &zero,
 841		.extra2		= &sixty,
 842	},
 843	{
 844		.procname	= "softlockup_panic",
 845		.data		= &softlockup_panic,
 846		.maxlen		= sizeof(int),
 847		.mode		= 0644,
 848		.proc_handler	= proc_dointvec_minmax,
 849		.extra1		= &zero,
 850		.extra2		= &one,
 851	},
 852	{
 853		.procname       = "nmi_watchdog",
 854		.data           = &watchdog_user_enabled,
 855		.maxlen         = sizeof (int),
 856		.mode           = 0644,
 857		.proc_handler   = proc_dowatchdog,
 858		.extra1		= &zero,
 859		.extra2		= &one,
 860	},
 861#endif
 862#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
 863	{
 864		.procname       = "unknown_nmi_panic",
 865		.data           = &unknown_nmi_panic,
 866		.maxlen         = sizeof (int),
 867		.mode           = 0644,
 868		.proc_handler   = proc_dointvec,
 869	},
 870#endif
 871#if defined(CONFIG_X86)
 872	{
 873		.procname	= "panic_on_unrecovered_nmi",
 874		.data		= &panic_on_unrecovered_nmi,
 875		.maxlen		= sizeof(int),
 876		.mode		= 0644,
 877		.proc_handler	= proc_dointvec,
 878	},
 879	{
 880		.procname	= "panic_on_io_nmi",
 881		.data		= &panic_on_io_nmi,
 882		.maxlen		= sizeof(int),
 883		.mode		= 0644,
 884		.proc_handler	= proc_dointvec,
 885	},
 886#ifdef CONFIG_DEBUG_STACKOVERFLOW
 887	{
 888		.procname	= "panic_on_stackoverflow",
 889		.data		= &sysctl_panic_on_stackoverflow,
 890		.maxlen		= sizeof(int),
 891		.mode		= 0644,
 892		.proc_handler	= proc_dointvec,
 893	},
 894#endif
 895	{
 896		.procname	= "bootloader_type",
 897		.data		= &bootloader_type,
 898		.maxlen		= sizeof (int),
 899		.mode		= 0444,
 900		.proc_handler	= proc_dointvec,
 901	},
 902	{
 903		.procname	= "bootloader_version",
 904		.data		= &bootloader_version,
 905		.maxlen		= sizeof (int),
 906		.mode		= 0444,
 907		.proc_handler	= proc_dointvec,
 908	},
 909	{
 910		.procname	= "kstack_depth_to_print",
 911		.data		= &kstack_depth_to_print,
 912		.maxlen		= sizeof(int),
 913		.mode		= 0644,
 914		.proc_handler	= proc_dointvec,
 915	},
 916	{
 917		.procname	= "io_delay_type",
 918		.data		= &io_delay_type,
 919		.maxlen		= sizeof(int),
 920		.mode		= 0644,
 921		.proc_handler	= proc_dointvec,
 922	},
 923#endif
 924#if defined(CONFIG_MMU)
 925	{
 926		.procname	= "randomize_va_space",
 927		.data		= &randomize_va_space,
 928		.maxlen		= sizeof(int),
 929		.mode		= 0644,
 930		.proc_handler	= proc_dointvec,
 931	},
 932#endif
 933#if defined(CONFIG_S390) && defined(CONFIG_SMP)
 934	{
 935		.procname	= "spin_retry",
 936		.data		= &spin_retry,
 937		.maxlen		= sizeof (int),
 938		.mode		= 0644,
 939		.proc_handler	= proc_dointvec,
 940	},
 941#endif
 942#if	defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
 943	{
 944		.procname	= "acpi_video_flags",
 945		.data		= &acpi_realmode_flags,
 946		.maxlen		= sizeof (unsigned long),
 947		.mode		= 0644,
 948		.proc_handler	= proc_doulongvec_minmax,
 949	},
 950#endif
 951#ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
 952	{
 953		.procname	= "ignore-unaligned-usertrap",
 954		.data		= &no_unaligned_warning,
 955		.maxlen		= sizeof (int),
 956	 	.mode		= 0644,
 957		.proc_handler	= proc_dointvec,
 958	},
 959#endif
 960#ifdef CONFIG_IA64
 961	{
 962		.procname	= "unaligned-dump-stack",
 963		.data		= &unaligned_dump_stack,
 964		.maxlen		= sizeof (int),
 965		.mode		= 0644,
 966		.proc_handler	= proc_dointvec,
 967	},
 968#endif
 969#ifdef CONFIG_DETECT_HUNG_TASK
 970	{
 971		.procname	= "hung_task_panic",
 972		.data		= &sysctl_hung_task_panic,
 973		.maxlen		= sizeof(int),
 974		.mode		= 0644,
 975		.proc_handler	= proc_dointvec_minmax,
 976		.extra1		= &zero,
 977		.extra2		= &one,
 978	},
 979	{
 980		.procname	= "hung_task_check_count",
 981		.data		= &sysctl_hung_task_check_count,
 982		.maxlen		= sizeof(int),
 983		.mode		= 0644,
 984		.proc_handler	= proc_dointvec_minmax,
 985		.extra1		= &zero,
 986	},
 987	{
 988		.procname	= "hung_task_timeout_secs",
 989		.data		= &sysctl_hung_task_timeout_secs,
 990		.maxlen		= sizeof(unsigned long),
 991		.mode		= 0644,
 992		.proc_handler	= proc_dohung_task_timeout_secs,
 993		.extra2		= &hung_task_timeout_max,
 994	},
 995	{
 996		.procname	= "hung_task_warnings",
 997		.data		= &sysctl_hung_task_warnings,
 998		.maxlen		= sizeof(int),
 999		.mode		= 0644,
1000		.proc_handler	= proc_dointvec_minmax,
1001		.extra1		= &neg_one,
1002	},
1003#endif
1004#ifdef CONFIG_COMPAT
1005	{
1006		.procname	= "compat-log",
1007		.data		= &compat_log,
1008		.maxlen		= sizeof (int),
1009	 	.mode		= 0644,
1010		.proc_handler	= proc_dointvec,
1011	},
1012#endif
1013#ifdef CONFIG_RT_MUTEXES
1014	{
1015		.procname	= "max_lock_depth",
1016		.data		= &max_lock_depth,
1017		.maxlen		= sizeof(int),
1018		.mode		= 0644,
1019		.proc_handler	= proc_dointvec,
1020	},
1021#endif
1022	{
1023		.procname	= "poweroff_cmd",
1024		.data		= &poweroff_cmd,
1025		.maxlen		= POWEROFF_CMD_PATH_LEN,
1026		.mode		= 0644,
1027		.proc_handler	= proc_dostring,
1028	},
1029#ifdef CONFIG_KEYS
1030	{
1031		.procname	= "keys",
1032		.mode		= 0555,
1033		.child		= key_sysctls,
1034	},
1035#endif
1036#ifdef CONFIG_RCU_TORTURE_TEST
1037	{
1038		.procname       = "rcutorture_runnable",
1039		.data           = &rcutorture_runnable,
1040		.maxlen         = sizeof(int),
1041		.mode           = 0644,
1042		.proc_handler	= proc_dointvec,
1043	},
1044#endif
1045#ifdef CONFIG_PERF_EVENTS
1046	/*
1047	 * User-space scripts rely on the existence of this file
1048	 * as a feature check for perf_events being enabled.
1049	 *
1050	 * So it's an ABI, do not remove!
1051	 */
1052	{
1053		.procname	= "perf_event_paranoid",
1054		.data		= &sysctl_perf_event_paranoid,
1055		.maxlen		= sizeof(sysctl_perf_event_paranoid),
1056		.mode		= 0644,
1057		.proc_handler	= proc_dointvec,
1058	},
1059	{
1060		.procname	= "perf_event_mlock_kb",
1061		.data		= &sysctl_perf_event_mlock,
1062		.maxlen		= sizeof(sysctl_perf_event_mlock),
1063		.mode		= 0644,
1064		.proc_handler	= proc_dointvec,
1065	},
1066	{
1067		.procname	= "perf_event_max_sample_rate",
1068		.data		= &sysctl_perf_event_sample_rate,
1069		.maxlen		= sizeof(sysctl_perf_event_sample_rate),
1070		.mode		= 0644,
1071		.proc_handler	= perf_proc_update_handler,
1072		.extra1		= &one,
1073	},
1074	{
1075		.procname	= "perf_cpu_time_max_percent",
1076		.data		= &sysctl_perf_cpu_time_max_percent,
1077		.maxlen		= sizeof(sysctl_perf_cpu_time_max_percent),
1078		.mode		= 0644,
1079		.proc_handler	= perf_cpu_time_max_percent_handler,
1080		.extra1		= &zero,
1081		.extra2		= &one_hundred,
1082	},
1083#endif
1084#ifdef CONFIG_KMEMCHECK
1085	{
1086		.procname	= "kmemcheck",
1087		.data		= &kmemcheck_enabled,
1088		.maxlen		= sizeof(int),
1089		.mode		= 0644,
1090		.proc_handler	= proc_dointvec,
1091	},
1092#endif
 
 
 
 
 
 
 
 
 
1093	{ }
1094};
1095
1096static struct ctl_table vm_table[] = {
1097	{
1098		.procname	= "overcommit_memory",
1099		.data		= &sysctl_overcommit_memory,
1100		.maxlen		= sizeof(sysctl_overcommit_memory),
1101		.mode		= 0644,
1102		.proc_handler	= proc_dointvec_minmax,
1103		.extra1		= &zero,
1104		.extra2		= &two,
1105	},
1106	{
1107		.procname	= "panic_on_oom",
1108		.data		= &sysctl_panic_on_oom,
1109		.maxlen		= sizeof(sysctl_panic_on_oom),
1110		.mode		= 0644,
1111		.proc_handler	= proc_dointvec_minmax,
1112		.extra1		= &zero,
1113		.extra2		= &two,
1114	},
1115	{
1116		.procname	= "oom_kill_allocating_task",
1117		.data		= &sysctl_oom_kill_allocating_task,
1118		.maxlen		= sizeof(sysctl_oom_kill_allocating_task),
1119		.mode		= 0644,
1120		.proc_handler	= proc_dointvec,
1121	},
1122	{
1123		.procname	= "oom_dump_tasks",
1124		.data		= &sysctl_oom_dump_tasks,
1125		.maxlen		= sizeof(sysctl_oom_dump_tasks),
1126		.mode		= 0644,
1127		.proc_handler	= proc_dointvec,
1128	},
1129	{
1130		.procname	= "overcommit_ratio",
1131		.data		= &sysctl_overcommit_ratio,
1132		.maxlen		= sizeof(sysctl_overcommit_ratio),
1133		.mode		= 0644,
1134		.proc_handler	= overcommit_ratio_handler,
1135	},
1136	{
1137		.procname	= "overcommit_kbytes",
1138		.data		= &sysctl_overcommit_kbytes,
1139		.maxlen		= sizeof(sysctl_overcommit_kbytes),
1140		.mode		= 0644,
1141		.proc_handler	= overcommit_kbytes_handler,
1142	},
1143	{
1144		.procname	= "page-cluster", 
1145		.data		= &page_cluster,
1146		.maxlen		= sizeof(int),
1147		.mode		= 0644,
1148		.proc_handler	= proc_dointvec_minmax,
1149		.extra1		= &zero,
1150	},
1151	{
1152		.procname	= "dirty_background_ratio",
1153		.data		= &dirty_background_ratio,
1154		.maxlen		= sizeof(dirty_background_ratio),
1155		.mode		= 0644,
1156		.proc_handler	= dirty_background_ratio_handler,
1157		.extra1		= &zero,
1158		.extra2		= &one_hundred,
1159	},
1160	{
1161		.procname	= "dirty_background_bytes",
1162		.data		= &dirty_background_bytes,
1163		.maxlen		= sizeof(dirty_background_bytes),
1164		.mode		= 0644,
1165		.proc_handler	= dirty_background_bytes_handler,
1166		.extra1		= &one_ul,
1167	},
1168	{
1169		.procname	= "dirty_ratio",
1170		.data		= &vm_dirty_ratio,
1171		.maxlen		= sizeof(vm_dirty_ratio),
1172		.mode		= 0644,
1173		.proc_handler	= dirty_ratio_handler,
1174		.extra1		= &zero,
1175		.extra2		= &one_hundred,
1176	},
1177	{
1178		.procname	= "dirty_bytes",
1179		.data		= &vm_dirty_bytes,
1180		.maxlen		= sizeof(vm_dirty_bytes),
1181		.mode		= 0644,
1182		.proc_handler	= dirty_bytes_handler,
1183		.extra1		= &dirty_bytes_min,
1184	},
1185	{
1186		.procname	= "dirty_writeback_centisecs",
1187		.data		= &dirty_writeback_interval,
1188		.maxlen		= sizeof(dirty_writeback_interval),
1189		.mode		= 0644,
1190		.proc_handler	= dirty_writeback_centisecs_handler,
1191	},
1192	{
1193		.procname	= "dirty_expire_centisecs",
1194		.data		= &dirty_expire_interval,
1195		.maxlen		= sizeof(dirty_expire_interval),
1196		.mode		= 0644,
1197		.proc_handler	= proc_dointvec_minmax,
1198		.extra1		= &zero,
1199	},
1200	{
1201		.procname       = "nr_pdflush_threads",
1202		.mode           = 0444 /* read-only */,
1203		.proc_handler   = pdflush_proc_obsolete,
 
 
1204	},
1205	{
1206		.procname	= "swappiness",
1207		.data		= &vm_swappiness,
1208		.maxlen		= sizeof(vm_swappiness),
1209		.mode		= 0644,
1210		.proc_handler	= proc_dointvec_minmax,
1211		.extra1		= &zero,
1212		.extra2		= &one_hundred,
1213	},
1214#ifdef CONFIG_HUGETLB_PAGE
1215	{
1216		.procname	= "nr_hugepages",
1217		.data		= NULL,
1218		.maxlen		= sizeof(unsigned long),
1219		.mode		= 0644,
1220		.proc_handler	= hugetlb_sysctl_handler,
1221		.extra1		= (void *)&hugetlb_zero,
1222		.extra2		= (void *)&hugetlb_infinity,
1223	},
1224#ifdef CONFIG_NUMA
1225	{
1226		.procname       = "nr_hugepages_mempolicy",
1227		.data           = NULL,
1228		.maxlen         = sizeof(unsigned long),
1229		.mode           = 0644,
1230		.proc_handler   = &hugetlb_mempolicy_sysctl_handler,
1231		.extra1		= (void *)&hugetlb_zero,
1232		.extra2		= (void *)&hugetlb_infinity,
1233	},
1234#endif
1235	 {
1236		.procname	= "hugetlb_shm_group",
1237		.data		= &sysctl_hugetlb_shm_group,
1238		.maxlen		= sizeof(gid_t),
1239		.mode		= 0644,
1240		.proc_handler	= proc_dointvec,
1241	 },
1242	 {
1243		.procname	= "hugepages_treat_as_movable",
1244		.data		= &hugepages_treat_as_movable,
1245		.maxlen		= sizeof(int),
1246		.mode		= 0644,
1247		.proc_handler	= proc_dointvec,
1248	},
1249	{
1250		.procname	= "nr_overcommit_hugepages",
1251		.data		= NULL,
1252		.maxlen		= sizeof(unsigned long),
1253		.mode		= 0644,
1254		.proc_handler	= hugetlb_overcommit_handler,
1255		.extra1		= (void *)&hugetlb_zero,
1256		.extra2		= (void *)&hugetlb_infinity,
1257	},
1258#endif
1259	{
1260		.procname	= "lowmem_reserve_ratio",
1261		.data		= &sysctl_lowmem_reserve_ratio,
1262		.maxlen		= sizeof(sysctl_lowmem_reserve_ratio),
1263		.mode		= 0644,
1264		.proc_handler	= lowmem_reserve_ratio_sysctl_handler,
1265	},
1266	{
1267		.procname	= "drop_caches",
1268		.data		= &sysctl_drop_caches,
1269		.maxlen		= sizeof(int),
1270		.mode		= 0644,
1271		.proc_handler	= drop_caches_sysctl_handler,
1272		.extra1		= &one,
1273		.extra2		= &four,
1274	},
1275#ifdef CONFIG_COMPACTION
1276	{
1277		.procname	= "compact_memory",
1278		.data		= &sysctl_compact_memory,
1279		.maxlen		= sizeof(int),
1280		.mode		= 0200,
1281		.proc_handler	= sysctl_compaction_handler,
1282	},
1283	{
1284		.procname	= "extfrag_threshold",
1285		.data		= &sysctl_extfrag_threshold,
1286		.maxlen		= sizeof(int),
1287		.mode		= 0644,
1288		.proc_handler	= sysctl_extfrag_handler,
1289		.extra1		= &min_extfrag_threshold,
1290		.extra2		= &max_extfrag_threshold,
1291	},
1292
1293#endif /* CONFIG_COMPACTION */
1294	{
1295		.procname	= "min_free_kbytes",
1296		.data		= &min_free_kbytes,
1297		.maxlen		= sizeof(min_free_kbytes),
1298		.mode		= 0644,
1299		.proc_handler	= min_free_kbytes_sysctl_handler,
1300		.extra1		= &zero,
1301	},
1302	{
1303		.procname	= "percpu_pagelist_fraction",
1304		.data		= &percpu_pagelist_fraction,
1305		.maxlen		= sizeof(percpu_pagelist_fraction),
1306		.mode		= 0644,
1307		.proc_handler	= percpu_pagelist_fraction_sysctl_handler,
1308		.extra1		= &min_percpu_pagelist_fract,
1309	},
1310#ifdef CONFIG_MMU
1311	{
1312		.procname	= "max_map_count",
1313		.data		= &sysctl_max_map_count,
1314		.maxlen		= sizeof(sysctl_max_map_count),
1315		.mode		= 0644,
1316		.proc_handler	= proc_dointvec_minmax,
1317		.extra1		= &zero,
1318	},
1319#else
1320	{
1321		.procname	= "nr_trim_pages",
1322		.data		= &sysctl_nr_trim_pages,
1323		.maxlen		= sizeof(sysctl_nr_trim_pages),
1324		.mode		= 0644,
1325		.proc_handler	= proc_dointvec_minmax,
1326		.extra1		= &zero,
1327	},
1328#endif
1329	{
1330		.procname	= "laptop_mode",
1331		.data		= &laptop_mode,
1332		.maxlen		= sizeof(laptop_mode),
1333		.mode		= 0644,
1334		.proc_handler	= proc_dointvec_jiffies,
1335	},
1336	{
1337		.procname	= "block_dump",
1338		.data		= &block_dump,
1339		.maxlen		= sizeof(block_dump),
1340		.mode		= 0644,
1341		.proc_handler	= proc_dointvec,
1342		.extra1		= &zero,
1343	},
1344	{
1345		.procname	= "vfs_cache_pressure",
1346		.data		= &sysctl_vfs_cache_pressure,
1347		.maxlen		= sizeof(sysctl_vfs_cache_pressure),
1348		.mode		= 0644,
1349		.proc_handler	= proc_dointvec,
1350		.extra1		= &zero,
1351	},
1352#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1353	{
1354		.procname	= "legacy_va_layout",
1355		.data		= &sysctl_legacy_va_layout,
1356		.maxlen		= sizeof(sysctl_legacy_va_layout),
1357		.mode		= 0644,
1358		.proc_handler	= proc_dointvec,
1359		.extra1		= &zero,
1360	},
1361#endif
1362#ifdef CONFIG_NUMA
1363	{
1364		.procname	= "zone_reclaim_mode",
1365		.data		= &zone_reclaim_mode,
1366		.maxlen		= sizeof(zone_reclaim_mode),
1367		.mode		= 0644,
1368		.proc_handler	= proc_dointvec,
1369		.extra1		= &zero,
1370	},
1371	{
1372		.procname	= "min_unmapped_ratio",
1373		.data		= &sysctl_min_unmapped_ratio,
1374		.maxlen		= sizeof(sysctl_min_unmapped_ratio),
1375		.mode		= 0644,
1376		.proc_handler	= sysctl_min_unmapped_ratio_sysctl_handler,
1377		.extra1		= &zero,
1378		.extra2		= &one_hundred,
1379	},
1380	{
1381		.procname	= "min_slab_ratio",
1382		.data		= &sysctl_min_slab_ratio,
1383		.maxlen		= sizeof(sysctl_min_slab_ratio),
1384		.mode		= 0644,
1385		.proc_handler	= sysctl_min_slab_ratio_sysctl_handler,
1386		.extra1		= &zero,
1387		.extra2		= &one_hundred,
1388	},
1389#endif
1390#ifdef CONFIG_SMP
1391	{
1392		.procname	= "stat_interval",
1393		.data		= &sysctl_stat_interval,
1394		.maxlen		= sizeof(sysctl_stat_interval),
1395		.mode		= 0644,
1396		.proc_handler	= proc_dointvec_jiffies,
1397	},
1398#endif
1399#ifdef CONFIG_MMU
1400	{
1401		.procname	= "mmap_min_addr",
1402		.data		= &dac_mmap_min_addr,
1403		.maxlen		= sizeof(unsigned long),
1404		.mode		= 0644,
1405		.proc_handler	= mmap_min_addr_handler,
1406	},
1407#endif
1408#ifdef CONFIG_NUMA
1409	{
1410		.procname	= "numa_zonelist_order",
1411		.data		= &numa_zonelist_order,
1412		.maxlen		= NUMA_ZONELIST_ORDER_LEN,
1413		.mode		= 0644,
1414		.proc_handler	= numa_zonelist_order_handler,
1415	},
1416#endif
1417#if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
1418   (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
1419	{
1420		.procname	= "vdso_enabled",
1421		.data		= &vdso_enabled,
1422		.maxlen		= sizeof(vdso_enabled),
1423		.mode		= 0644,
1424		.proc_handler	= proc_dointvec,
1425		.extra1		= &zero,
1426	},
1427#endif
1428#ifdef CONFIG_HIGHMEM
1429	{
1430		.procname	= "highmem_is_dirtyable",
1431		.data		= &vm_highmem_is_dirtyable,
1432		.maxlen		= sizeof(vm_highmem_is_dirtyable),
1433		.mode		= 0644,
1434		.proc_handler	= proc_dointvec_minmax,
1435		.extra1		= &zero,
1436		.extra2		= &one,
1437	},
1438#endif
1439	{
1440		.procname	= "scan_unevictable_pages",
1441		.data		= &scan_unevictable_pages,
1442		.maxlen		= sizeof(scan_unevictable_pages),
1443		.mode		= 0644,
1444		.proc_handler	= scan_unevictable_handler,
1445	},
1446#ifdef CONFIG_MEMORY_FAILURE
1447	{
1448		.procname	= "memory_failure_early_kill",
1449		.data		= &sysctl_memory_failure_early_kill,
1450		.maxlen		= sizeof(sysctl_memory_failure_early_kill),
1451		.mode		= 0644,
1452		.proc_handler	= proc_dointvec_minmax,
1453		.extra1		= &zero,
1454		.extra2		= &one,
1455	},
1456	{
1457		.procname	= "memory_failure_recovery",
1458		.data		= &sysctl_memory_failure_recovery,
1459		.maxlen		= sizeof(sysctl_memory_failure_recovery),
1460		.mode		= 0644,
1461		.proc_handler	= proc_dointvec_minmax,
1462		.extra1		= &zero,
1463		.extra2		= &one,
1464	},
1465#endif
1466	{
1467		.procname	= "user_reserve_kbytes",
1468		.data		= &sysctl_user_reserve_kbytes,
1469		.maxlen		= sizeof(sysctl_user_reserve_kbytes),
1470		.mode		= 0644,
1471		.proc_handler	= proc_doulongvec_minmax,
1472	},
1473	{
1474		.procname	= "admin_reserve_kbytes",
1475		.data		= &sysctl_admin_reserve_kbytes,
1476		.maxlen		= sizeof(sysctl_admin_reserve_kbytes),
1477		.mode		= 0644,
1478		.proc_handler	= proc_doulongvec_minmax,
1479	},
1480	{ }
1481};
1482
1483#if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1484static struct ctl_table binfmt_misc_table[] = {
1485	{ }
1486};
1487#endif
1488
1489static struct ctl_table fs_table[] = {
1490	{
1491		.procname	= "inode-nr",
1492		.data		= &inodes_stat,
1493		.maxlen		= 2*sizeof(long),
1494		.mode		= 0444,
1495		.proc_handler	= proc_nr_inodes,
1496	},
1497	{
1498		.procname	= "inode-state",
1499		.data		= &inodes_stat,
1500		.maxlen		= 7*sizeof(long),
1501		.mode		= 0444,
1502		.proc_handler	= proc_nr_inodes,
1503	},
1504	{
1505		.procname	= "file-nr",
1506		.data		= &files_stat,
1507		.maxlen		= sizeof(files_stat),
1508		.mode		= 0444,
1509		.proc_handler	= proc_nr_files,
1510	},
1511	{
1512		.procname	= "file-max",
1513		.data		= &files_stat.max_files,
1514		.maxlen		= sizeof(files_stat.max_files),
1515		.mode		= 0644,
1516		.proc_handler	= proc_doulongvec_minmax,
1517	},
1518	{
1519		.procname	= "nr_open",
1520		.data		= &sysctl_nr_open,
1521		.maxlen		= sizeof(int),
1522		.mode		= 0644,
1523		.proc_handler	= proc_dointvec_minmax,
1524		.extra1		= &sysctl_nr_open_min,
1525		.extra2		= &sysctl_nr_open_max,
1526	},
1527	{
1528		.procname	= "dentry-state",
1529		.data		= &dentry_stat,
1530		.maxlen		= 6*sizeof(long),
1531		.mode		= 0444,
1532		.proc_handler	= proc_nr_dentry,
1533	},
1534	{
1535		.procname	= "overflowuid",
1536		.data		= &fs_overflowuid,
1537		.maxlen		= sizeof(int),
1538		.mode		= 0644,
1539		.proc_handler	= proc_dointvec_minmax,
1540		.extra1		= &minolduid,
1541		.extra2		= &maxolduid,
1542	},
1543	{
1544		.procname	= "overflowgid",
1545		.data		= &fs_overflowgid,
1546		.maxlen		= sizeof(int),
1547		.mode		= 0644,
1548		.proc_handler	= proc_dointvec_minmax,
1549		.extra1		= &minolduid,
1550		.extra2		= &maxolduid,
1551	},
1552#ifdef CONFIG_FILE_LOCKING
1553	{
1554		.procname	= "leases-enable",
1555		.data		= &leases_enable,
1556		.maxlen		= sizeof(int),
1557		.mode		= 0644,
1558		.proc_handler	= proc_dointvec,
1559	},
1560#endif
1561#ifdef CONFIG_DNOTIFY
1562	{
1563		.procname	= "dir-notify-enable",
1564		.data		= &dir_notify_enable,
1565		.maxlen		= sizeof(int),
1566		.mode		= 0644,
1567		.proc_handler	= proc_dointvec,
1568	},
1569#endif
1570#ifdef CONFIG_MMU
1571#ifdef CONFIG_FILE_LOCKING
1572	{
1573		.procname	= "lease-break-time",
1574		.data		= &lease_break_time,
1575		.maxlen		= sizeof(int),
1576		.mode		= 0644,
1577		.proc_handler	= proc_dointvec,
1578	},
1579#endif
1580#ifdef CONFIG_AIO
1581	{
1582		.procname	= "aio-nr",
1583		.data		= &aio_nr,
1584		.maxlen		= sizeof(aio_nr),
1585		.mode		= 0444,
1586		.proc_handler	= proc_doulongvec_minmax,
1587	},
1588	{
1589		.procname	= "aio-max-nr",
1590		.data		= &aio_max_nr,
1591		.maxlen		= sizeof(aio_max_nr),
1592		.mode		= 0644,
1593		.proc_handler	= proc_doulongvec_minmax,
1594	},
1595#endif /* CONFIG_AIO */
1596#ifdef CONFIG_INOTIFY_USER
1597	{
1598		.procname	= "inotify",
1599		.mode		= 0555,
1600		.child		= inotify_table,
1601	},
1602#endif	
1603#ifdef CONFIG_EPOLL
1604	{
1605		.procname	= "epoll",
1606		.mode		= 0555,
1607		.child		= epoll_table,
1608	},
1609#endif
1610#endif
1611	{
1612		.procname	= "protected_symlinks",
1613		.data		= &sysctl_protected_symlinks,
1614		.maxlen		= sizeof(int),
1615		.mode		= 0600,
1616		.proc_handler	= proc_dointvec_minmax,
1617		.extra1		= &zero,
1618		.extra2		= &one,
1619	},
1620	{
1621		.procname	= "protected_hardlinks",
1622		.data		= &sysctl_protected_hardlinks,
1623		.maxlen		= sizeof(int),
1624		.mode		= 0600,
1625		.proc_handler	= proc_dointvec_minmax,
1626		.extra1		= &zero,
1627		.extra2		= &one,
1628	},
1629	{
1630		.procname	= "suid_dumpable",
1631		.data		= &suid_dumpable,
1632		.maxlen		= sizeof(int),
1633		.mode		= 0644,
1634		.proc_handler	= proc_dointvec_minmax_coredump,
1635		.extra1		= &zero,
1636		.extra2		= &two,
1637	},
1638#if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1639	{
1640		.procname	= "binfmt_misc",
1641		.mode		= 0555,
1642		.child		= binfmt_misc_table,
1643	},
1644#endif
1645	{
1646		.procname	= "pipe-max-size",
1647		.data		= &pipe_max_size,
1648		.maxlen		= sizeof(int),
1649		.mode		= 0644,
1650		.proc_handler	= &pipe_proc_fn,
1651		.extra1		= &pipe_min_size,
1652	},
1653	{ }
1654};
1655
1656static struct ctl_table debug_table[] = {
1657#ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
 
1658	{
1659		.procname	= "exception-trace",
1660		.data		= &show_unhandled_signals,
1661		.maxlen		= sizeof(int),
1662		.mode		= 0644,
1663		.proc_handler	= proc_dointvec
1664	},
1665#endif
1666#if defined(CONFIG_OPTPROBES)
1667	{
1668		.procname	= "kprobes-optimization",
1669		.data		= &sysctl_kprobes_optimization,
1670		.maxlen		= sizeof(int),
1671		.mode		= 0644,
1672		.proc_handler	= proc_kprobes_optimization_handler,
1673		.extra1		= &zero,
1674		.extra2		= &one,
1675	},
1676#endif
1677	{ }
1678};
1679
1680static struct ctl_table dev_table[] = {
1681	{ }
1682};
1683
1684int __init sysctl_init(void)
1685{
1686	struct ctl_table_header *hdr;
1687
1688	hdr = register_sysctl_table(sysctl_base_table);
1689	kmemleak_not_leak(hdr);
1690	return 0;
1691}
1692
1693#endif /* CONFIG_SYSCTL */
1694
1695/*
1696 * /proc/sys support
1697 */
1698
1699#ifdef CONFIG_PROC_SYSCTL
1700
1701static int _proc_do_string(void* data, int maxlen, int write,
1702			   void __user *buffer,
1703			   size_t *lenp, loff_t *ppos)
1704{
1705	size_t len;
1706	char __user *p;
1707	char c;
1708
1709	if (!data || !maxlen || !*lenp) {
1710		*lenp = 0;
1711		return 0;
1712	}
1713
1714	if (write) {
1715		len = 0;
1716		p = buffer;
1717		while (len < *lenp) {
1718			if (get_user(c, p++))
1719				return -EFAULT;
1720			if (c == 0 || c == '\n')
1721				break;
1722			len++;
1723		}
1724		if (len >= maxlen)
1725			len = maxlen-1;
1726		if(copy_from_user(data, buffer, len))
1727			return -EFAULT;
1728		((char *) data)[len] = 0;
1729		*ppos += *lenp;
1730	} else {
1731		len = strlen(data);
1732		if (len > maxlen)
1733			len = maxlen;
1734
1735		if (*ppos > len) {
1736			*lenp = 0;
1737			return 0;
1738		}
1739
1740		data += *ppos;
1741		len  -= *ppos;
1742
1743		if (len > *lenp)
1744			len = *lenp;
1745		if (len)
1746			if(copy_to_user(buffer, data, len))
1747				return -EFAULT;
1748		if (len < *lenp) {
1749			if(put_user('\n', ((char __user *) buffer) + len))
1750				return -EFAULT;
1751			len++;
1752		}
1753		*lenp = len;
1754		*ppos += len;
1755	}
1756	return 0;
1757}
1758
1759/**
1760 * proc_dostring - read a string sysctl
1761 * @table: the sysctl table
1762 * @write: %TRUE if this is a write to the sysctl file
1763 * @buffer: the user buffer
1764 * @lenp: the size of the user buffer
1765 * @ppos: file position
1766 *
1767 * Reads/writes a string from/to the user buffer. If the kernel
1768 * buffer provided is not large enough to hold the string, the
1769 * string is truncated. The copied string is %NULL-terminated.
1770 * If the string is being read by the user process, it is copied
1771 * and a newline '\n' is added. It is truncated if the buffer is
1772 * not large enough.
1773 *
1774 * Returns 0 on success.
1775 */
1776int proc_dostring(struct ctl_table *table, int write,
1777		  void __user *buffer, size_t *lenp, loff_t *ppos)
1778{
1779	return _proc_do_string(table->data, table->maxlen, write,
1780			       buffer, lenp, ppos);
1781}
1782
1783static size_t proc_skip_spaces(char **buf)
1784{
1785	size_t ret;
1786	char *tmp = skip_spaces(*buf);
1787	ret = tmp - *buf;
1788	*buf = tmp;
1789	return ret;
1790}
1791
1792static void proc_skip_char(char **buf, size_t *size, const char v)
1793{
1794	while (*size) {
1795		if (**buf != v)
1796			break;
1797		(*size)--;
1798		(*buf)++;
1799	}
1800}
1801
1802#define TMPBUFLEN 22
1803/**
1804 * proc_get_long - reads an ASCII formatted integer from a user buffer
1805 *
1806 * @buf: a kernel buffer
1807 * @size: size of the kernel buffer
1808 * @val: this is where the number will be stored
1809 * @neg: set to %TRUE if number is negative
1810 * @perm_tr: a vector which contains the allowed trailers
1811 * @perm_tr_len: size of the perm_tr vector
1812 * @tr: pointer to store the trailer character
1813 *
1814 * In case of success %0 is returned and @buf and @size are updated with
1815 * the amount of bytes read. If @tr is non-NULL and a trailing
1816 * character exists (size is non-zero after returning from this
1817 * function), @tr is updated with the trailing character.
1818 */
1819static int proc_get_long(char **buf, size_t *size,
1820			  unsigned long *val, bool *neg,
1821			  const char *perm_tr, unsigned perm_tr_len, char *tr)
1822{
1823	int len;
1824	char *p, tmp[TMPBUFLEN];
1825
1826	if (!*size)
1827		return -EINVAL;
1828
1829	len = *size;
1830	if (len > TMPBUFLEN - 1)
1831		len = TMPBUFLEN - 1;
1832
1833	memcpy(tmp, *buf, len);
1834
1835	tmp[len] = 0;
1836	p = tmp;
1837	if (*p == '-' && *size > 1) {
1838		*neg = true;
1839		p++;
1840	} else
1841		*neg = false;
1842	if (!isdigit(*p))
1843		return -EINVAL;
1844
1845	*val = simple_strtoul(p, &p, 0);
1846
1847	len = p - tmp;
1848
1849	/* We don't know if the next char is whitespace thus we may accept
1850	 * invalid integers (e.g. 1234...a) or two integers instead of one
1851	 * (e.g. 123...1). So lets not allow such large numbers. */
1852	if (len == TMPBUFLEN - 1)
1853		return -EINVAL;
1854
1855	if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
1856		return -EINVAL;
1857
1858	if (tr && (len < *size))
1859		*tr = *p;
1860
1861	*buf += len;
1862	*size -= len;
1863
1864	return 0;
1865}
1866
1867/**
1868 * proc_put_long - converts an integer to a decimal ASCII formatted string
1869 *
1870 * @buf: the user buffer
1871 * @size: the size of the user buffer
1872 * @val: the integer to be converted
1873 * @neg: sign of the number, %TRUE for negative
1874 *
1875 * In case of success %0 is returned and @buf and @size are updated with
1876 * the amount of bytes written.
1877 */
1878static int proc_put_long(void __user **buf, size_t *size, unsigned long val,
1879			  bool neg)
1880{
1881	int len;
1882	char tmp[TMPBUFLEN], *p = tmp;
1883
1884	sprintf(p, "%s%lu", neg ? "-" : "", val);
1885	len = strlen(tmp);
1886	if (len > *size)
1887		len = *size;
1888	if (copy_to_user(*buf, tmp, len))
1889		return -EFAULT;
1890	*size -= len;
1891	*buf += len;
1892	return 0;
1893}
1894#undef TMPBUFLEN
1895
1896static int proc_put_char(void __user **buf, size_t *size, char c)
1897{
1898	if (*size) {
1899		char __user **buffer = (char __user **)buf;
1900		if (put_user(c, *buffer))
1901			return -EFAULT;
1902		(*size)--, (*buffer)++;
1903		*buf = *buffer;
1904	}
1905	return 0;
1906}
1907
1908static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
1909				 int *valp,
1910				 int write, void *data)
1911{
1912	if (write) {
1913		*valp = *negp ? -*lvalp : *lvalp;
1914	} else {
1915		int val = *valp;
1916		if (val < 0) {
1917			*negp = true;
1918			*lvalp = (unsigned long)-val;
1919		} else {
1920			*negp = false;
1921			*lvalp = (unsigned long)val;
1922		}
1923	}
1924	return 0;
1925}
1926
1927static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
1928
1929static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
1930		  int write, void __user *buffer,
1931		  size_t *lenp, loff_t *ppos,
1932		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
1933			      int write, void *data),
1934		  void *data)
1935{
1936	int *i, vleft, first = 1, err = 0;
1937	unsigned long page = 0;
1938	size_t left;
1939	char *kbuf;
1940	
1941	if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
1942		*lenp = 0;
1943		return 0;
1944	}
1945	
1946	i = (int *) tbl_data;
1947	vleft = table->maxlen / sizeof(*i);
1948	left = *lenp;
1949
1950	if (!conv)
1951		conv = do_proc_dointvec_conv;
1952
1953	if (write) {
1954		if (left > PAGE_SIZE - 1)
1955			left = PAGE_SIZE - 1;
1956		page = __get_free_page(GFP_TEMPORARY);
1957		kbuf = (char *) page;
1958		if (!kbuf)
1959			return -ENOMEM;
1960		if (copy_from_user(kbuf, buffer, left)) {
1961			err = -EFAULT;
1962			goto free;
1963		}
1964		kbuf[left] = 0;
1965	}
1966
1967	for (; left && vleft--; i++, first=0) {
1968		unsigned long lval;
1969		bool neg;
1970
1971		if (write) {
1972			left -= proc_skip_spaces(&kbuf);
1973
1974			if (!left)
1975				break;
1976			err = proc_get_long(&kbuf, &left, &lval, &neg,
1977					     proc_wspace_sep,
1978					     sizeof(proc_wspace_sep), NULL);
1979			if (err)
1980				break;
1981			if (conv(&neg, &lval, i, 1, data)) {
1982				err = -EINVAL;
1983				break;
1984			}
1985		} else {
1986			if (conv(&neg, &lval, i, 0, data)) {
1987				err = -EINVAL;
1988				break;
1989			}
1990			if (!first)
1991				err = proc_put_char(&buffer, &left, '\t');
1992			if (err)
1993				break;
1994			err = proc_put_long(&buffer, &left, lval, neg);
1995			if (err)
1996				break;
1997		}
1998	}
1999
2000	if (!write && !first && left && !err)
2001		err = proc_put_char(&buffer, &left, '\n');
2002	if (write && !err && left)
2003		left -= proc_skip_spaces(&kbuf);
2004free:
2005	if (write) {
2006		free_page(page);
2007		if (first)
2008			return err ? : -EINVAL;
2009	}
2010	*lenp -= left;
2011	*ppos += *lenp;
2012	return err;
2013}
2014
2015static int do_proc_dointvec(struct ctl_table *table, int write,
2016		  void __user *buffer, size_t *lenp, loff_t *ppos,
2017		  int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
2018			      int write, void *data),
2019		  void *data)
2020{
2021	return __do_proc_dointvec(table->data, table, write,
2022			buffer, lenp, ppos, conv, data);
2023}
2024
2025/**
2026 * proc_dointvec - read a vector of integers
2027 * @table: the sysctl table
2028 * @write: %TRUE if this is a write to the sysctl file
2029 * @buffer: the user buffer
2030 * @lenp: the size of the user buffer
2031 * @ppos: file position
2032 *
2033 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2034 * values from/to the user buffer, treated as an ASCII string. 
2035 *
2036 * Returns 0 on success.
2037 */
2038int proc_dointvec(struct ctl_table *table, int write,
2039		     void __user *buffer, size_t *lenp, loff_t *ppos)
2040{
2041    return do_proc_dointvec(table,write,buffer,lenp,ppos,
2042		    	    NULL,NULL);
2043}
2044
2045/*
2046 * Taint values can only be increased
2047 * This means we can safely use a temporary.
2048 */
2049static int proc_taint(struct ctl_table *table, int write,
2050			       void __user *buffer, size_t *lenp, loff_t *ppos)
2051{
2052	struct ctl_table t;
2053	unsigned long tmptaint = get_taint();
2054	int err;
2055
2056	if (write && !capable(CAP_SYS_ADMIN))
2057		return -EPERM;
2058
2059	t = *table;
2060	t.data = &tmptaint;
2061	err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
2062	if (err < 0)
2063		return err;
2064
2065	if (write) {
2066		/*
2067		 * Poor man's atomic or. Not worth adding a primitive
2068		 * to everyone's atomic.h for this
2069		 */
2070		int i;
2071		for (i = 0; i < BITS_PER_LONG && tmptaint >> i; i++) {
2072			if ((tmptaint >> i) & 1)
2073				add_taint(i, LOCKDEP_STILL_OK);
2074		}
2075	}
2076
2077	return err;
2078}
2079
2080#ifdef CONFIG_PRINTK
2081static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
2082				void __user *buffer, size_t *lenp, loff_t *ppos)
2083{
2084	if (write && !capable(CAP_SYS_ADMIN))
2085		return -EPERM;
2086
2087	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2088}
2089#endif
2090
2091struct do_proc_dointvec_minmax_conv_param {
2092	int *min;
2093	int *max;
2094};
2095
2096static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
2097					int *valp,
2098					int write, void *data)
2099{
2100	struct do_proc_dointvec_minmax_conv_param *param = data;
2101	if (write) {
2102		int val = *negp ? -*lvalp : *lvalp;
2103		if ((param->min && *param->min > val) ||
2104		    (param->max && *param->max < val))
2105			return -EINVAL;
2106		*valp = val;
2107	} else {
2108		int val = *valp;
2109		if (val < 0) {
2110			*negp = true;
2111			*lvalp = (unsigned long)-val;
2112		} else {
2113			*negp = false;
2114			*lvalp = (unsigned long)val;
2115		}
2116	}
2117	return 0;
2118}
2119
2120/**
2121 * proc_dointvec_minmax - read a vector of integers with min/max values
2122 * @table: the sysctl table
2123 * @write: %TRUE if this is a write to the sysctl file
2124 * @buffer: the user buffer
2125 * @lenp: the size of the user buffer
2126 * @ppos: file position
2127 *
2128 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2129 * values from/to the user buffer, treated as an ASCII string.
2130 *
2131 * This routine will ensure the values are within the range specified by
2132 * table->extra1 (min) and table->extra2 (max).
2133 *
2134 * Returns 0 on success.
2135 */
2136int proc_dointvec_minmax(struct ctl_table *table, int write,
2137		  void __user *buffer, size_t *lenp, loff_t *ppos)
2138{
2139	struct do_proc_dointvec_minmax_conv_param param = {
2140		.min = (int *) table->extra1,
2141		.max = (int *) table->extra2,
2142	};
2143	return do_proc_dointvec(table, write, buffer, lenp, ppos,
2144				do_proc_dointvec_minmax_conv, &param);
2145}
2146
2147static void validate_coredump_safety(void)
2148{
2149#ifdef CONFIG_COREDUMP
2150	if (suid_dumpable == SUID_DUMP_ROOT &&
2151	    core_pattern[0] != '/' && core_pattern[0] != '|') {
2152		printk(KERN_WARNING "Unsafe core_pattern used with "\
2153			"suid_dumpable=2. Pipe handler or fully qualified "\
2154			"core dump path required.\n");
2155	}
2156#endif
2157}
2158
2159static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
2160		void __user *buffer, size_t *lenp, loff_t *ppos)
2161{
2162	int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2163	if (!error)
2164		validate_coredump_safety();
2165	return error;
2166}
2167
2168#ifdef CONFIG_COREDUMP
2169static int proc_dostring_coredump(struct ctl_table *table, int write,
2170		  void __user *buffer, size_t *lenp, loff_t *ppos)
2171{
2172	int error = proc_dostring(table, write, buffer, lenp, ppos);
2173	if (!error)
2174		validate_coredump_safety();
2175	return error;
2176}
2177#endif
2178
2179static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table, int write,
2180				     void __user *buffer,
2181				     size_t *lenp, loff_t *ppos,
2182				     unsigned long convmul,
2183				     unsigned long convdiv)
2184{
2185	unsigned long *i, *min, *max;
2186	int vleft, first = 1, err = 0;
2187	unsigned long page = 0;
2188	size_t left;
2189	char *kbuf;
2190
2191	if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
2192		*lenp = 0;
2193		return 0;
2194	}
2195
2196	i = (unsigned long *) data;
2197	min = (unsigned long *) table->extra1;
2198	max = (unsigned long *) table->extra2;
2199	vleft = table->maxlen / sizeof(unsigned long);
2200	left = *lenp;
2201
2202	if (write) {
2203		if (left > PAGE_SIZE - 1)
2204			left = PAGE_SIZE - 1;
2205		page = __get_free_page(GFP_TEMPORARY);
2206		kbuf = (char *) page;
2207		if (!kbuf)
2208			return -ENOMEM;
2209		if (copy_from_user(kbuf, buffer, left)) {
2210			err = -EFAULT;
2211			goto free;
2212		}
2213		kbuf[left] = 0;
2214	}
2215
2216	for (; left && vleft--; i++, first = 0) {
2217		unsigned long val;
2218
2219		if (write) {
2220			bool neg;
2221
2222			left -= proc_skip_spaces(&kbuf);
2223
2224			err = proc_get_long(&kbuf, &left, &val, &neg,
2225					     proc_wspace_sep,
2226					     sizeof(proc_wspace_sep), NULL);
2227			if (err)
2228				break;
2229			if (neg)
2230				continue;
2231			if ((min && val < *min) || (max && val > *max))
2232				continue;
2233			*i = val;
2234		} else {
2235			val = convdiv * (*i) / convmul;
2236			if (!first) {
2237				err = proc_put_char(&buffer, &left, '\t');
2238				if (err)
2239					break;
2240			}
2241			err = proc_put_long(&buffer, &left, val, false);
2242			if (err)
2243				break;
2244		}
2245	}
2246
2247	if (!write && !first && left && !err)
2248		err = proc_put_char(&buffer, &left, '\n');
2249	if (write && !err)
2250		left -= proc_skip_spaces(&kbuf);
2251free:
2252	if (write) {
2253		free_page(page);
2254		if (first)
2255			return err ? : -EINVAL;
2256	}
2257	*lenp -= left;
2258	*ppos += *lenp;
2259	return err;
2260}
2261
2262static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
2263				     void __user *buffer,
2264				     size_t *lenp, loff_t *ppos,
2265				     unsigned long convmul,
2266				     unsigned long convdiv)
2267{
2268	return __do_proc_doulongvec_minmax(table->data, table, write,
2269			buffer, lenp, ppos, convmul, convdiv);
2270}
2271
2272/**
2273 * proc_doulongvec_minmax - read a vector of long integers with min/max values
2274 * @table: the sysctl table
2275 * @write: %TRUE if this is a write to the sysctl file
2276 * @buffer: the user buffer
2277 * @lenp: the size of the user buffer
2278 * @ppos: file position
2279 *
2280 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2281 * values from/to the user buffer, treated as an ASCII string.
2282 *
2283 * This routine will ensure the values are within the range specified by
2284 * table->extra1 (min) and table->extra2 (max).
2285 *
2286 * Returns 0 on success.
2287 */
2288int proc_doulongvec_minmax(struct ctl_table *table, int write,
2289			   void __user *buffer, size_t *lenp, loff_t *ppos)
2290{
2291    return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
2292}
2293
2294/**
2295 * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
2296 * @table: the sysctl table
2297 * @write: %TRUE if this is a write to the sysctl file
2298 * @buffer: the user buffer
2299 * @lenp: the size of the user buffer
2300 * @ppos: file position
2301 *
2302 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2303 * values from/to the user buffer, treated as an ASCII string. The values
2304 * are treated as milliseconds, and converted to jiffies when they are stored.
2305 *
2306 * This routine will ensure the values are within the range specified by
2307 * table->extra1 (min) and table->extra2 (max).
2308 *
2309 * Returns 0 on success.
2310 */
2311int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
2312				      void __user *buffer,
2313				      size_t *lenp, loff_t *ppos)
2314{
2315    return do_proc_doulongvec_minmax(table, write, buffer,
2316				     lenp, ppos, HZ, 1000l);
2317}
2318
2319
2320static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
2321					 int *valp,
2322					 int write, void *data)
2323{
2324	if (write) {
2325		if (*lvalp > LONG_MAX / HZ)
2326			return 1;
2327		*valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
2328	} else {
2329		int val = *valp;
2330		unsigned long lval;
2331		if (val < 0) {
2332			*negp = true;
2333			lval = (unsigned long)-val;
2334		} else {
2335			*negp = false;
2336			lval = (unsigned long)val;
2337		}
2338		*lvalp = lval / HZ;
2339	}
2340	return 0;
2341}
2342
2343static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
2344						int *valp,
2345						int write, void *data)
2346{
2347	if (write) {
2348		if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
2349			return 1;
2350		*valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
2351	} else {
2352		int val = *valp;
2353		unsigned long lval;
2354		if (val < 0) {
2355			*negp = true;
2356			lval = (unsigned long)-val;
2357		} else {
2358			*negp = false;
2359			lval = (unsigned long)val;
2360		}
2361		*lvalp = jiffies_to_clock_t(lval);
2362	}
2363	return 0;
2364}
2365
2366static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
2367					    int *valp,
2368					    int write, void *data)
2369{
2370	if (write) {
2371		unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
2372
2373		if (jif > INT_MAX)
2374			return 1;
2375		*valp = (int)jif;
2376	} else {
2377		int val = *valp;
2378		unsigned long lval;
2379		if (val < 0) {
2380			*negp = true;
2381			lval = (unsigned long)-val;
2382		} else {
2383			*negp = false;
2384			lval = (unsigned long)val;
2385		}
2386		*lvalp = jiffies_to_msecs(lval);
2387	}
2388	return 0;
2389}
2390
2391/**
2392 * proc_dointvec_jiffies - read a vector of integers as seconds
2393 * @table: the sysctl table
2394 * @write: %TRUE if this is a write to the sysctl file
2395 * @buffer: the user buffer
2396 * @lenp: the size of the user buffer
2397 * @ppos: file position
2398 *
2399 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2400 * values from/to the user buffer, treated as an ASCII string. 
2401 * The values read are assumed to be in seconds, and are converted into
2402 * jiffies.
2403 *
2404 * Returns 0 on success.
2405 */
2406int proc_dointvec_jiffies(struct ctl_table *table, int write,
2407			  void __user *buffer, size_t *lenp, loff_t *ppos)
2408{
2409    return do_proc_dointvec(table,write,buffer,lenp,ppos,
2410		    	    do_proc_dointvec_jiffies_conv,NULL);
2411}
2412
2413/**
2414 * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
2415 * @table: the sysctl table
2416 * @write: %TRUE if this is a write to the sysctl file
2417 * @buffer: the user buffer
2418 * @lenp: the size of the user buffer
2419 * @ppos: pointer to the file position
2420 *
2421 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2422 * values from/to the user buffer, treated as an ASCII string. 
2423 * The values read are assumed to be in 1/USER_HZ seconds, and 
2424 * are converted into jiffies.
2425 *
2426 * Returns 0 on success.
2427 */
2428int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
2429				 void __user *buffer, size_t *lenp, loff_t *ppos)
2430{
2431    return do_proc_dointvec(table,write,buffer,lenp,ppos,
2432		    	    do_proc_dointvec_userhz_jiffies_conv,NULL);
2433}
2434
2435/**
2436 * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
2437 * @table: the sysctl table
2438 * @write: %TRUE if this is a write to the sysctl file
2439 * @buffer: the user buffer
2440 * @lenp: the size of the user buffer
2441 * @ppos: file position
2442 * @ppos: the current position in the file
2443 *
2444 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2445 * values from/to the user buffer, treated as an ASCII string. 
2446 * The values read are assumed to be in 1/1000 seconds, and 
2447 * are converted into jiffies.
2448 *
2449 * Returns 0 on success.
2450 */
2451int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
2452			     void __user *buffer, size_t *lenp, loff_t *ppos)
2453{
2454	return do_proc_dointvec(table, write, buffer, lenp, ppos,
2455				do_proc_dointvec_ms_jiffies_conv, NULL);
2456}
2457
2458static int proc_do_cad_pid(struct ctl_table *table, int write,
2459			   void __user *buffer, size_t *lenp, loff_t *ppos)
2460{
2461	struct pid *new_pid;
2462	pid_t tmp;
2463	int r;
2464
2465	tmp = pid_vnr(cad_pid);
2466
2467	r = __do_proc_dointvec(&tmp, table, write, buffer,
2468			       lenp, ppos, NULL, NULL);
2469	if (r || !write)
2470		return r;
2471
2472	new_pid = find_get_pid(tmp);
2473	if (!new_pid)
2474		return -ESRCH;
2475
2476	put_pid(xchg(&cad_pid, new_pid));
2477	return 0;
2478}
2479
2480/**
2481 * proc_do_large_bitmap - read/write from/to a large bitmap
2482 * @table: the sysctl table
2483 * @write: %TRUE if this is a write to the sysctl file
2484 * @buffer: the user buffer
2485 * @lenp: the size of the user buffer
2486 * @ppos: file position
2487 *
2488 * The bitmap is stored at table->data and the bitmap length (in bits)
2489 * in table->maxlen.
2490 *
2491 * We use a range comma separated format (e.g. 1,3-4,10-10) so that
2492 * large bitmaps may be represented in a compact manner. Writing into
2493 * the file will clear the bitmap then update it with the given input.
2494 *
2495 * Returns 0 on success.
2496 */
2497int proc_do_large_bitmap(struct ctl_table *table, int write,
2498			 void __user *buffer, size_t *lenp, loff_t *ppos)
2499{
2500	int err = 0;
2501	bool first = 1;
2502	size_t left = *lenp;
2503	unsigned long bitmap_len = table->maxlen;
2504	unsigned long *bitmap = (unsigned long *) table->data;
2505	unsigned long *tmp_bitmap = NULL;
2506	char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
2507
2508	if (!bitmap_len || !left || (*ppos && !write)) {
2509		*lenp = 0;
2510		return 0;
2511	}
2512
2513	if (write) {
2514		unsigned long page = 0;
2515		char *kbuf;
2516
2517		if (left > PAGE_SIZE - 1)
2518			left = PAGE_SIZE - 1;
2519
2520		page = __get_free_page(GFP_TEMPORARY);
2521		kbuf = (char *) page;
2522		if (!kbuf)
2523			return -ENOMEM;
2524		if (copy_from_user(kbuf, buffer, left)) {
2525			free_page(page);
2526			return -EFAULT;
2527                }
2528		kbuf[left] = 0;
2529
2530		tmp_bitmap = kzalloc(BITS_TO_LONGS(bitmap_len) * sizeof(unsigned long),
2531				     GFP_KERNEL);
2532		if (!tmp_bitmap) {
2533			free_page(page);
2534			return -ENOMEM;
2535		}
2536		proc_skip_char(&kbuf, &left, '\n');
2537		while (!err && left) {
2538			unsigned long val_a, val_b;
2539			bool neg;
2540
2541			err = proc_get_long(&kbuf, &left, &val_a, &neg, tr_a,
2542					     sizeof(tr_a), &c);
2543			if (err)
2544				break;
2545			if (val_a >= bitmap_len || neg) {
2546				err = -EINVAL;
2547				break;
2548			}
2549
2550			val_b = val_a;
2551			if (left) {
2552				kbuf++;
2553				left--;
2554			}
2555
2556			if (c == '-') {
2557				err = proc_get_long(&kbuf, &left, &val_b,
2558						     &neg, tr_b, sizeof(tr_b),
2559						     &c);
2560				if (err)
2561					break;
2562				if (val_b >= bitmap_len || neg ||
2563				    val_a > val_b) {
2564					err = -EINVAL;
2565					break;
2566				}
2567				if (left) {
2568					kbuf++;
2569					left--;
2570				}
2571			}
2572
2573			bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
2574			first = 0;
2575			proc_skip_char(&kbuf, &left, '\n');
2576		}
2577		free_page(page);
2578	} else {
2579		unsigned long bit_a, bit_b = 0;
2580
2581		while (left) {
2582			bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
2583			if (bit_a >= bitmap_len)
2584				break;
2585			bit_b = find_next_zero_bit(bitmap, bitmap_len,
2586						   bit_a + 1) - 1;
2587
2588			if (!first) {
2589				err = proc_put_char(&buffer, &left, ',');
2590				if (err)
2591					break;
2592			}
2593			err = proc_put_long(&buffer, &left, bit_a, false);
2594			if (err)
2595				break;
2596			if (bit_a != bit_b) {
2597				err = proc_put_char(&buffer, &left, '-');
2598				if (err)
2599					break;
2600				err = proc_put_long(&buffer, &left, bit_b, false);
2601				if (err)
2602					break;
2603			}
2604
2605			first = 0; bit_b++;
2606		}
2607		if (!err)
2608			err = proc_put_char(&buffer, &left, '\n');
2609	}
2610
2611	if (!err) {
2612		if (write) {
2613			if (*ppos)
2614				bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
2615			else
2616				bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
2617		}
2618		kfree(tmp_bitmap);
2619		*lenp -= left;
2620		*ppos += *lenp;
2621		return 0;
2622	} else {
2623		kfree(tmp_bitmap);
2624		return err;
2625	}
2626}
2627
2628#else /* CONFIG_PROC_SYSCTL */
2629
2630int proc_dostring(struct ctl_table *table, int write,
2631		  void __user *buffer, size_t *lenp, loff_t *ppos)
2632{
2633	return -ENOSYS;
2634}
2635
2636int proc_dointvec(struct ctl_table *table, int write,
2637		  void __user *buffer, size_t *lenp, loff_t *ppos)
2638{
2639	return -ENOSYS;
2640}
2641
2642int proc_dointvec_minmax(struct ctl_table *table, int write,
2643		    void __user *buffer, size_t *lenp, loff_t *ppos)
2644{
2645	return -ENOSYS;
2646}
2647
2648int proc_dointvec_jiffies(struct ctl_table *table, int write,
2649		    void __user *buffer, size_t *lenp, loff_t *ppos)
2650{
2651	return -ENOSYS;
2652}
2653
2654int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
2655		    void __user *buffer, size_t *lenp, loff_t *ppos)
2656{
2657	return -ENOSYS;
2658}
2659
2660int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
2661			     void __user *buffer, size_t *lenp, loff_t *ppos)
2662{
2663	return -ENOSYS;
2664}
2665
2666int proc_doulongvec_minmax(struct ctl_table *table, int write,
2667		    void __user *buffer, size_t *lenp, loff_t *ppos)
2668{
2669	return -ENOSYS;
2670}
2671
2672int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
2673				      void __user *buffer,
2674				      size_t *lenp, loff_t *ppos)
2675{
2676    return -ENOSYS;
2677}
2678
2679
2680#endif /* CONFIG_PROC_SYSCTL */
2681
2682/*
2683 * No sense putting this after each symbol definition, twice,
2684 * exception granted :-)
2685 */
2686EXPORT_SYMBOL(proc_dointvec);
2687EXPORT_SYMBOL(proc_dointvec_jiffies);
2688EXPORT_SYMBOL(proc_dointvec_minmax);
2689EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
2690EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
2691EXPORT_SYMBOL(proc_dostring);
2692EXPORT_SYMBOL(proc_doulongvec_minmax);
2693EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);