projects / GitHub / moto-9609 / android_kernel_motorola_exynos9610.git / blob
? search:


9d75510813cf9e8a4d2d2fa6c40a8de4d3215630
[GitHub/moto-9609/android_kernel_motorola_exynos9610.git] / kernel / sysctl.c
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/aio.h>
23 #include <linux/mm.h>
24 #include <linux/swap.h>
25 #include <linux/slab.h>
26 #include <linux/sysctl.h>
27 #include <linux/bitmap.h>
28 #include <linux/signal.h>
29 #include <linux/printk.h>
30 #include <linux/proc_fs.h>
31 #include <linux/security.h>
32 #include <linux/ctype.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/sched/coredump.h>
66 #include <linux/kexec.h>
67 #include <linux/bpf.h>
68 #include <linux/mount.h>
69
70 #include <linux/uaccess.h>
71 #include <asm/processor.h>
72
73 #ifdef CONFIG_X86
74 #include <asm/nmi.h>
75 #include <asm/stacktrace.h>
76 #include <asm/io.h>
77 #endif
78 #ifdef CONFIG_SPARC
79 #include <asm/setup.h>
80 #endif
81 #ifdef CONFIG_BSD_PROCESS_ACCT
82 #include <linux/acct.h>
83 #endif
84 #ifdef CONFIG_RT_MUTEXES
85 #include <linux/rtmutex.h>
86 #endif
87 #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
88 #include <linux/lockdep.h>
89 #endif
90 #ifdef CONFIG_CHR_DEV_SG
91 #include <scsi/sg.h>
92 #endif
93
94 #ifdef CONFIG_LOCKUP_DETECTOR
95 #include <linux/nmi.h>
96 #endif
97
98 #if defined(CONFIG_SYSCTL)
99
100 /* External variables not in a header file. */
101 extern int suid_dumpable;
102 #ifdef CONFIG_COREDUMP
103 extern int core_uses_pid;
104 extern char core_pattern[];
105 extern unsigned int core_pipe_limit;
106 #endif
107 extern int pid_max;
108 extern int extra_free_kbytes;
109 extern int pid_max_min, pid_max_max;
110 extern int percpu_pagelist_fraction;
111 extern int latencytop_enabled;
112 extern unsigned int sysctl_nr_open_min, sysctl_nr_open_max;
113 #ifndef CONFIG_MMU
114 extern int sysctl_nr_trim_pages;
115 #endif
116
117 /* Constants used for minimum and maximum */
118 #ifdef CONFIG_LOCKUP_DETECTOR
119 static int sixty = 60;
120 #endif
121
122 static int __maybe_unused neg_one = -1;
123
124 static int zero;
125 static int __maybe_unused one = 1;
126 static int __maybe_unused two = 2;
127 static int __maybe_unused four = 4;
128 static unsigned long one_ul = 1;
129 static int one_hundred = 100;
130 static int one_thousand = 1000;
131 #ifdef CONFIG_PRINTK
132 static int ten_thousand = 10000;
133 #endif
134 #ifdef CONFIG_PERF_EVENTS
135 static int six_hundred_forty_kb = 640 * 1024;
136 #endif
137
138 /* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */
139 static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;
140
141 /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
142 static int maxolduid = 65535;
143 static int minolduid;
144
145 static int ngroups_max = NGROUPS_MAX;
146 static const int cap_last_cap = CAP_LAST_CAP;
147
148 /*this is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs */
149 #ifdef CONFIG_DETECT_HUNG_TASK
150 static unsigned long hung_task_timeout_max = (LONG_MAX/HZ);
151 #endif
152
153 #ifdef CONFIG_INOTIFY_USER
154 #include <linux/inotify.h>
155 #endif
156 #ifdef CONFIG_SPARC
157 #endif
158
159 #ifdef __hppa__
160 extern int pwrsw_enabled;
161 #endif
162
163 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
164 extern int unaligned_enabled;
165 #endif
166
167 #ifdef CONFIG_IA64
168 extern int unaligned_dump_stack;
169 #endif
170
171 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
172 extern int no_unaligned_warning;
173 #endif
174
175 #ifdef CONFIG_PROC_SYSCTL
176
177 /**
178 * enum sysctl_writes_mode - supported sysctl write modes
179 *
180 * @SYSCTL_WRITES_LEGACY: each write syscall must fully contain the sysctl value
181 * to be written, and multiple writes on the same sysctl file descriptor
182 * will rewrite the sysctl value, regardless of file position. No warning
183 * is issued when the initial position is not 0.
184 * @SYSCTL_WRITES_WARN: same as above but warn when the initial file position is
185 * not 0.
186 * @SYSCTL_WRITES_STRICT: writes to numeric sysctl entries must always be at
187 * file position 0 and the value must be fully contained in the buffer
188 * sent to the write syscall. If dealing with strings respect the file
189 * position, but restrict this to the max length of the buffer, anything
190 * passed the max lenght will be ignored. Multiple writes will append
191 * to the buffer.
192 *
193 * These write modes control how current file position affects the behavior of
194 * updating sysctl values through the proc interface on each write.
195 */
196 enum sysctl_writes_mode {
197 SYSCTL_WRITES_LEGACY = -1,
198 SYSCTL_WRITES_WARN = 0,
199 SYSCTL_WRITES_STRICT = 1,
200 };
201
202 static enum sysctl_writes_mode sysctl_writes_strict = SYSCTL_WRITES_STRICT;
203
204 static int proc_do_cad_pid(struct ctl_table *table, int write,
205 void __user *buffer, size_t *lenp, loff_t *ppos);
206 static int proc_taint(struct ctl_table *table, int write,
207 void __user *buffer, size_t *lenp, loff_t *ppos);
208 #endif
209
210 #ifdef CONFIG_PRINTK
211 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
212 void __user *buffer, size_t *lenp, loff_t *ppos);
213 #endif
214
215 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
216 void __user *buffer, size_t *lenp, loff_t *ppos);
217 #ifdef CONFIG_COREDUMP
218 static int proc_dostring_coredump(struct ctl_table *table, int write,
219 void __user *buffer, size_t *lenp, loff_t *ppos);
220 #endif
221
222 #ifdef CONFIG_MAGIC_SYSRQ
223 /* Note: sysrq code uses it's own private copy */
224 static int __sysrq_enabled = CONFIG_MAGIC_SYSRQ_DEFAULT_ENABLE;
225
226 static int sysrq_sysctl_handler(struct ctl_table *table, int write,
227 void __user *buffer, size_t *lenp,
228 loff_t *ppos)
229 {
230 int error;
231
232 error = proc_dointvec(table, write, buffer, lenp, ppos);
233 if (error)
234 return error;
235
236 if (write)
237 sysrq_toggle_support(__sysrq_enabled);
238
239 return 0;
240 }
241
242 #endif
243
244 static struct ctl_table kern_table[];
245 static struct ctl_table vm_table[];
246 static struct ctl_table fs_table[];
247 static struct ctl_table debug_table[];
248 static struct ctl_table dev_table[];
249 extern struct ctl_table random_table[];
250 #ifdef CONFIG_EPOLL
251 extern struct ctl_table epoll_table[];
252 #endif
253
254 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
255 int sysctl_legacy_va_layout;
256 #endif
257
258 /* The default sysctl tables: */
259
260 static struct ctl_table sysctl_base_table[] = {
261 {
262 .procname = "kernel",
263 .mode = 0555,
264 .child = kern_table,
265 },
266 {
267 .procname = "vm",
268 .mode = 0555,
269 .child = vm_table,
270 },
271 {
272 .procname = "fs",
273 .mode = 0555,
274 .child = fs_table,
275 },
276 {
277 .procname = "debug",
278 .mode = 0555,
279 .child = debug_table,
280 },
281 {
282 .procname = "dev",
283 .mode = 0555,
284 .child = dev_table,
285 },
286 { }
287 };
288
289 #ifdef CONFIG_SCHED_DEBUG
290 static int min_sched_granularity_ns = 100000; /* 100 usecs */
291 static int max_sched_granularity_ns = NSEC_PER_SEC; /* 1 second */
292 static int min_wakeup_granularity_ns; /* 0 usecs */
293 static int max_wakeup_granularity_ns = NSEC_PER_SEC; /* 1 second */
294 #ifdef CONFIG_SMP
295 static int min_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
296 static int max_sched_tunable_scaling = SCHED_TUNABLESCALING_END-1;
297 #endif /* CONFIG_SMP */
298 #endif /* CONFIG_SCHED_DEBUG */
299
300 #ifdef CONFIG_COMPACTION
301 static int min_extfrag_threshold;
302 static int max_extfrag_threshold = 1000;
303 #endif
304
305 static struct ctl_table kern_table[] = {
306 {
307 .procname = "sched_child_runs_first",
308 .data = &sysctl_sched_child_runs_first,
309 .maxlen = sizeof(unsigned int),
310 .mode = 0644,
311 .proc_handler = proc_dointvec,
312 },
313 #ifdef CONFIG_SCHED_DEBUG
314 {
315 .procname = "sched_min_granularity_ns",
316 .data = &sysctl_sched_min_granularity,
317 .maxlen = sizeof(unsigned int),
318 .mode = 0644,
319 .proc_handler = sched_proc_update_handler,
320 .extra1 = &min_sched_granularity_ns,
321 .extra2 = &max_sched_granularity_ns,
322 },
323 {
324 .procname = "sched_latency_ns",
325 .data = &sysctl_sched_latency,
326 .maxlen = sizeof(unsigned int),
327 .mode = 0644,
328 .proc_handler = sched_proc_update_handler,
329 .extra1 = &min_sched_granularity_ns,
330 .extra2 = &max_sched_granularity_ns,
331 },
332 #ifdef CONFIG_SCHED_WALT
333 {
334 .procname = "sched_use_walt_cpu_util",
335 .data = &sysctl_sched_use_walt_cpu_util,
336 .maxlen = sizeof(unsigned int),
337 .mode = 0644,
338 .proc_handler = proc_dointvec,
339 },
340 {
341 .procname = "sched_use_walt_task_util",
342 .data = &sysctl_sched_use_walt_task_util,
343 .maxlen = sizeof(unsigned int),
344 .mode = 0644,
345 .proc_handler = proc_dointvec,
346 },
347 {
348 .procname = "sched_walt_init_task_load_pct",
349 .data = &sysctl_sched_walt_init_task_load_pct,
350 .maxlen = sizeof(unsigned int),
351 .mode = 0644,
352 .proc_handler = proc_dointvec,
353 },
354 {
355 .procname = "sched_walt_cpu_high_irqload",
356 .data = &sysctl_sched_walt_cpu_high_irqload,
357 .maxlen = sizeof(unsigned int),
358 .mode = 0644,
359 .proc_handler = proc_dointvec,
360 },
361 #endif
362 {
363 .procname = "sched_sync_hint_enable",
364 .data = &sysctl_sched_sync_hint_enable,
365 .maxlen = sizeof(unsigned int),
366 .mode = 0644,
367 .proc_handler = proc_dointvec,
368 },
369 {
370 .procname = "sched_cstate_aware",
371 .data = &sysctl_sched_cstate_aware,
372 .maxlen = sizeof(unsigned int),
373 .mode = 0644,
374 .proc_handler = proc_dointvec,
375 },
376 {
377 .procname = "sched_wakeup_granularity_ns",
378 .data = &sysctl_sched_wakeup_granularity,
379 .maxlen = sizeof(unsigned int),
380 .mode = 0644,
381 .proc_handler = sched_proc_update_handler,
382 .extra1 = &min_wakeup_granularity_ns,
383 .extra2 = &max_wakeup_granularity_ns,
384 },
385 #ifdef CONFIG_SMP
386 {
387 .procname = "sched_tunable_scaling",
388 .data = &sysctl_sched_tunable_scaling,
389 .maxlen = sizeof(enum sched_tunable_scaling),
390 .mode = 0644,
391 .proc_handler = sched_proc_update_handler,
392 .extra1 = &min_sched_tunable_scaling,
393 .extra2 = &max_sched_tunable_scaling,
394 },
395 {
396 .procname = "sched_migration_cost_ns",
397 .data = &sysctl_sched_migration_cost,
398 .maxlen = sizeof(unsigned int),
399 .mode = 0644,
400 .proc_handler = proc_dointvec,
401 },
402 {
403 .procname = "sched_nr_migrate",
404 .data = &sysctl_sched_nr_migrate,
405 .maxlen = sizeof(unsigned int),
406 .mode = 0644,
407 .proc_handler = proc_dointvec,
408 },
409 {
410 .procname = "sched_time_avg_ms",
411 .data = &sysctl_sched_time_avg,
412 .maxlen = sizeof(unsigned int),
413 .mode = 0644,
414 .proc_handler = proc_dointvec_minmax,
415 .extra1 = &one,
416 },
417 #ifdef CONFIG_SCHEDSTATS
418 {
419 .procname = "sched_schedstats",
420 .data = NULL,
421 .maxlen = sizeof(unsigned int),
422 .mode = 0644,
423 .proc_handler = sysctl_schedstats,
424 .extra1 = &zero,
425 .extra2 = &one,
426 },
427 #endif /* CONFIG_SCHEDSTATS */
428 #endif /* CONFIG_SMP */
429 #ifdef CONFIG_NUMA_BALANCING
430 {
431 .procname = "numa_balancing_scan_delay_ms",
432 .data = &sysctl_numa_balancing_scan_delay,
433 .maxlen = sizeof(unsigned int),
434 .mode = 0644,
435 .proc_handler = proc_dointvec,
436 },
437 {
438 .procname = "numa_balancing_scan_period_min_ms",
439 .data = &sysctl_numa_balancing_scan_period_min,
440 .maxlen = sizeof(unsigned int),
441 .mode = 0644,
442 .proc_handler = proc_dointvec,
443 },
444 {
445 .procname = "numa_balancing_scan_period_max_ms",
446 .data = &sysctl_numa_balancing_scan_period_max,
447 .maxlen = sizeof(unsigned int),
448 .mode = 0644,
449 .proc_handler = proc_dointvec,
450 },
451 {
452 .procname = "numa_balancing_scan_size_mb",
453 .data = &sysctl_numa_balancing_scan_size,
454 .maxlen = sizeof(unsigned int),
455 .mode = 0644,
456 .proc_handler = proc_dointvec_minmax,
457 .extra1 = &one,
458 },
459 {
460 .procname = "numa_balancing",
461 .data = NULL, /* filled in by handler */
462 .maxlen = sizeof(unsigned int),
463 .mode = 0644,
464 .proc_handler = sysctl_numa_balancing,
465 .extra1 = &zero,
466 .extra2 = &one,
467 },
468 #endif /* CONFIG_NUMA_BALANCING */
469 #endif /* CONFIG_SCHED_DEBUG */
470 {
471 .procname = "sched_rt_period_us",
472 .data = &sysctl_sched_rt_period,
473 .maxlen = sizeof(unsigned int),
474 .mode = 0644,
475 .proc_handler = sched_rt_handler,
476 },
477 {
478 .procname = "sched_rt_runtime_us",
479 .data = &sysctl_sched_rt_runtime,
480 .maxlen = sizeof(int),
481 .mode = 0644,
482 .proc_handler = sched_rt_handler,
483 },
484 {
485 .procname = "sched_rr_timeslice_ms",
486 .data = &sysctl_sched_rr_timeslice,
487 .maxlen = sizeof(int),
488 .mode = 0644,
489 .proc_handler = sched_rr_handler,
490 },
491 #ifdef CONFIG_SCHED_AUTOGROUP
492 {
493 .procname = "sched_autogroup_enabled",
494 .data = &sysctl_sched_autogroup_enabled,
495 .maxlen = sizeof(unsigned int),
496 .mode = 0644,
497 .proc_handler = proc_dointvec_minmax,
498 .extra1 = &zero,
499 .extra2 = &one,
500 },
501 #endif
502 #ifdef CONFIG_CFS_BANDWIDTH
503 {
504 .procname = "sched_cfs_bandwidth_slice_us",
505 .data = &sysctl_sched_cfs_bandwidth_slice,
506 .maxlen = sizeof(unsigned int),
507 .mode = 0644,
508 .proc_handler = proc_dointvec_minmax,
509 .extra1 = &one,
510 },
511 #endif
512 #ifdef CONFIG_PROVE_LOCKING
513 {
514 .procname = "prove_locking",
515 .data = &prove_locking,
516 .maxlen = sizeof(int),
517 .mode = 0644,
518 .proc_handler = proc_dointvec,
519 },
520 #endif
521 #ifdef CONFIG_LOCK_STAT
522 {
523 .procname = "lock_stat",
524 .data = &lock_stat,
525 .maxlen = sizeof(int),
526 .mode = 0644,
527 .proc_handler = proc_dointvec,
528 },
529 #endif
530 {
531 .procname = "panic",
532 .data = &panic_timeout,
533 .maxlen = sizeof(int),
534 .mode = 0644,
535 .proc_handler = proc_dointvec,
536 },
537 #ifdef CONFIG_COREDUMP
538 {
539 .procname = "core_uses_pid",
540 .data = &core_uses_pid,
541 .maxlen = sizeof(int),
542 .mode = 0644,
543 .proc_handler = proc_dointvec,
544 },
545 {
546 .procname = "core_pattern",
547 .data = core_pattern,
548 .maxlen = CORENAME_MAX_SIZE,
549 .mode = 0644,
550 .proc_handler = proc_dostring_coredump,
551 },
552 {
553 .procname = "core_pipe_limit",
554 .data = &core_pipe_limit,
555 .maxlen = sizeof(unsigned int),
556 .mode = 0644,
557 .proc_handler = proc_dointvec,
558 },
559 #endif
560 #ifdef CONFIG_PROC_SYSCTL
561 {
562 .procname = "tainted",
563 .maxlen = sizeof(long),
564 .mode = 0644,
565 .proc_handler = proc_taint,
566 },
567 {
568 .procname = "sysctl_writes_strict",
569 .data = &sysctl_writes_strict,
570 .maxlen = sizeof(int),
571 .mode = 0644,
572 .proc_handler = proc_dointvec_minmax,
573 .extra1 = &neg_one,
574 .extra2 = &one,
575 },
576 #endif
577 #ifdef CONFIG_LATENCYTOP
578 {
579 .procname = "latencytop",
580 .data = &latencytop_enabled,
581 .maxlen = sizeof(int),
582 .mode = 0644,
583 .proc_handler = sysctl_latencytop,
584 },
585 #endif
586 #ifdef CONFIG_BLK_DEV_INITRD
587 {
588 .procname = "real-root-dev",
589 .data = &real_root_dev,
590 .maxlen = sizeof(int),
591 .mode = 0644,
592 .proc_handler = proc_dointvec,
593 },
594 #endif
595 {
596 .procname = "print-fatal-signals",
597 .data = &print_fatal_signals,
598 .maxlen = sizeof(int),
599 .mode = 0644,
600 .proc_handler = proc_dointvec,
601 },
602 #ifdef CONFIG_SPARC
603 {
604 .procname = "reboot-cmd",
605 .data = reboot_command,
606 .maxlen = 256,
607 .mode = 0644,
608 .proc_handler = proc_dostring,
609 },
610 {
611 .procname = "stop-a",
612 .data = &stop_a_enabled,
613 .maxlen = sizeof (int),
614 .mode = 0644,
615 .proc_handler = proc_dointvec,
616 },
617 {
618 .procname = "scons-poweroff",
619 .data = &scons_pwroff,
620 .maxlen = sizeof (int),
621 .mode = 0644,
622 .proc_handler = proc_dointvec,
623 },
624 #endif
625 #ifdef CONFIG_SPARC64
626 {
627 .procname = "tsb-ratio",
628 .data = &sysctl_tsb_ratio,
629 .maxlen = sizeof (int),
630 .mode = 0644,
631 .proc_handler = proc_dointvec,
632 },
633 #endif
634 #ifdef __hppa__
635 {
636 .procname = "soft-power",
637 .data = &pwrsw_enabled,
638 .maxlen = sizeof (int),
639 .mode = 0644,
640 .proc_handler = proc_dointvec,
641 },
642 #endif
643 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW
644 {
645 .procname = "unaligned-trap",
646 .data = &unaligned_enabled,
647 .maxlen = sizeof (int),
648 .mode = 0644,
649 .proc_handler = proc_dointvec,
650 },
651 #endif
652 {
653 .procname = "ctrl-alt-del",
654 .data = &C_A_D,
655 .maxlen = sizeof(int),
656 .mode = 0644,
657 .proc_handler = proc_dointvec,
658 },
659 #ifdef CONFIG_FUNCTION_TRACER
660 {
661 .procname = "ftrace_enabled",
662 .data = &ftrace_enabled,
663 .maxlen = sizeof(int),
664 .mode = 0644,
665 .proc_handler = ftrace_enable_sysctl,
666 },
667 #endif
668 #ifdef CONFIG_STACK_TRACER
669 {
670 .procname = "stack_tracer_enabled",
671 .data = &stack_tracer_enabled,
672 .maxlen = sizeof(int),
673 .mode = 0644,
674 .proc_handler = stack_trace_sysctl,
675 },
676 #endif
677 #ifdef CONFIG_TRACING
678 {
679 .procname = "ftrace_dump_on_oops",
680 .data = &ftrace_dump_on_oops,
681 .maxlen = sizeof(int),
682 .mode = 0644,
683 .proc_handler = proc_dointvec,
684 },
685 {
686 .procname = "traceoff_on_warning",
687 .data = &__disable_trace_on_warning,
688 .maxlen = sizeof(__disable_trace_on_warning),
689 .mode = 0644,
690 .proc_handler = proc_dointvec,
691 },
692 {
693 .procname = "tracepoint_printk",
694 .data = &tracepoint_printk,
695 .maxlen = sizeof(tracepoint_printk),
696 .mode = 0644,
697 .proc_handler = tracepoint_printk_sysctl,
698 },
699 #endif
700 #ifdef CONFIG_KEXEC_CORE
701 {
702 .procname = "kexec_load_disabled",
703 .data = &kexec_load_disabled,
704 .maxlen = sizeof(int),
705 .mode = 0644,
706 /* only handle a transition from default "0" to "1" */
707 .proc_handler = proc_dointvec_minmax,
708 .extra1 = &one,
709 .extra2 = &one,
710 },
711 #endif
712 #ifdef CONFIG_MODULES
713 {
714 .procname = "modprobe",
715 .data = &modprobe_path,
716 .maxlen = KMOD_PATH_LEN,
717 .mode = 0644,
718 .proc_handler = proc_dostring,
719 },
720 {
721 .procname = "modules_disabled",
722 .data = &modules_disabled,
723 .maxlen = sizeof(int),
724 .mode = 0644,
725 /* only handle a transition from default "0" to "1" */
726 .proc_handler = proc_dointvec_minmax,
727 .extra1 = &one,
728 .extra2 = &one,
729 },
730 #endif
731 #ifdef CONFIG_UEVENT_HELPER
732 {
733 .procname = "hotplug",
734 .data = &uevent_helper,
735 .maxlen = UEVENT_HELPER_PATH_LEN,
736 .mode = 0644,
737 .proc_handler = proc_dostring,
738 },
739 #endif
740 #ifdef CONFIG_CHR_DEV_SG
741 {
742 .procname = "sg-big-buff",
743 .data = &sg_big_buff,
744 .maxlen = sizeof (int),
745 .mode = 0444,
746 .proc_handler = proc_dointvec,
747 },
748 #endif
749 #ifdef CONFIG_BSD_PROCESS_ACCT
750 {
751 .procname = "acct",
752 .data = &acct_parm,
753 .maxlen = 3*sizeof(int),
754 .mode = 0644,
755 .proc_handler = proc_dointvec,
756 },
757 #endif
758 #ifdef CONFIG_MAGIC_SYSRQ
759 {
760 .procname = "sysrq",
761 .data = &__sysrq_enabled,
762 .maxlen = sizeof (int),
763 .mode = 0644,
764 .proc_handler = sysrq_sysctl_handler,
765 },
766 #endif
767 #ifdef CONFIG_PROC_SYSCTL
768 {
769 .procname = "cad_pid",
770 .data = NULL,
771 .maxlen = sizeof (int),
772 .mode = 0600,
773 .proc_handler = proc_do_cad_pid,
774 },
775 #endif
776 {
777 .procname = "threads-max",
778 .data = NULL,
779 .maxlen = sizeof(int),
780 .mode = 0644,
781 .proc_handler = sysctl_max_threads,
782 },
783 {
784 .procname = "random",
785 .mode = 0555,
786 .child = random_table,
787 },
788 {
789 .procname = "usermodehelper",
790 .mode = 0555,
791 .child = usermodehelper_table,
792 },
793 {
794 .procname = "overflowuid",
795 .data = &overflowuid,
796 .maxlen = sizeof(int),
797 .mode = 0644,
798 .proc_handler = proc_dointvec_minmax,
799 .extra1 = &minolduid,
800 .extra2 = &maxolduid,
801 },
802 {
803 .procname = "overflowgid",
804 .data = &overflowgid,
805 .maxlen = sizeof(int),
806 .mode = 0644,
807 .proc_handler = proc_dointvec_minmax,
808 .extra1 = &minolduid,
809 .extra2 = &maxolduid,
810 },
811 #ifdef CONFIG_S390
812 #ifdef CONFIG_MATHEMU
813 {
814 .procname = "ieee_emulation_warnings",
815 .data = &sysctl_ieee_emulation_warnings,
816 .maxlen = sizeof(int),
817 .mode = 0644,
818 .proc_handler = proc_dointvec,
819 },
820 #endif
821 {
822 .procname = "userprocess_debug",
823 .data = &show_unhandled_signals,
824 .maxlen = sizeof(int),
825 .mode = 0644,
826 .proc_handler = proc_dointvec,
827 },
828 #endif
829 {
830 .procname = "pid_max",
831 .data = &pid_max,
832 .maxlen = sizeof (int),
833 .mode = 0644,
834 .proc_handler = proc_dointvec_minmax,
835 .extra1 = &pid_max_min,
836 .extra2 = &pid_max_max,
837 },
838 {
839 .procname = "panic_on_oops",
840 .data = &panic_on_oops,
841 .maxlen = sizeof(int),
842 .mode = 0644,
843 .proc_handler = proc_dointvec,
844 },
845 #if defined CONFIG_PRINTK
846 {
847 .procname = "printk",
848 .data = &console_loglevel,
849 .maxlen = 4*sizeof(int),
850 .mode = 0644,
851 .proc_handler = proc_dointvec,
852 },
853 {
854 .procname = "printk_ratelimit",
855 .data = &printk_ratelimit_state.interval,
856 .maxlen = sizeof(int),
857 .mode = 0644,
858 .proc_handler = proc_dointvec_jiffies,
859 },
860 {
861 .procname = "printk_ratelimit_burst",
862 .data = &printk_ratelimit_state.burst,
863 .maxlen = sizeof(int),
864 .mode = 0644,
865 .proc_handler = proc_dointvec,
866 },
867 {
868 .procname = "printk_delay",
869 .data = &printk_delay_msec,
870 .maxlen = sizeof(int),
871 .mode = 0644,
872 .proc_handler = proc_dointvec_minmax,
873 .extra1 = &zero,
874 .extra2 = &ten_thousand,
875 },
876 {
877 .procname = "printk_devkmsg",
878 .data = devkmsg_log_str,
879 .maxlen = DEVKMSG_STR_MAX_SIZE,
880 .mode = 0644,
881 .proc_handler = devkmsg_sysctl_set_loglvl,
882 },
883 {
884 .procname = "dmesg_restrict",
885 .data = &dmesg_restrict,
886 .maxlen = sizeof(int),
887 .mode = 0644,
888 .proc_handler = proc_dointvec_minmax_sysadmin,
889 .extra1 = &zero,
890 .extra2 = &one,
891 },
892 {
893 .procname = "kptr_restrict",
894 .data = &kptr_restrict,
895 .maxlen = sizeof(int),
896 .mode = 0644,
897 .proc_handler = proc_dointvec_minmax_sysadmin,
898 .extra1 = &zero,
899 .extra2 = &two,
900 },
901 #endif
902 {
903 .procname = "ngroups_max",
904 .data = &ngroups_max,
905 .maxlen = sizeof (int),
906 .mode = 0444,
907 .proc_handler = proc_dointvec,
908 },
909 {
910 .procname = "cap_last_cap",
911 .data = (void *)&cap_last_cap,
912 .maxlen = sizeof(int),
913 .mode = 0444,
914 .proc_handler = proc_dointvec,
915 },
916 #if defined(CONFIG_LOCKUP_DETECTOR)
917 {
918 .procname = "watchdog",
919 .data = &watchdog_user_enabled,
920 .maxlen = sizeof(int),
921 .mode = 0644,
922 .proc_handler = proc_watchdog,
923 .extra1 = &zero,
924 .extra2 = &one,
925 },
926 {
927 .procname = "watchdog_thresh",
928 .data = &watchdog_thresh,
929 .maxlen = sizeof(int),
930 .mode = 0644,
931 .proc_handler = proc_watchdog_thresh,
932 .extra1 = &zero,
933 .extra2 = &sixty,
934 },
935 {
936 .procname = "nmi_watchdog",
937 .data = &nmi_watchdog_user_enabled,
938 .maxlen = sizeof(int),
939 .mode = NMI_WATCHDOG_SYSCTL_PERM,
940 .proc_handler = proc_nmi_watchdog,
941 .extra1 = &zero,
942 .extra2 = &one,
943 },
944 {
945 .procname = "watchdog_cpumask",
946 .data = &watchdog_cpumask_bits,
947 .maxlen = NR_CPUS,
948 .mode = 0644,
949 .proc_handler = proc_watchdog_cpumask,
950 },
951 #ifdef CONFIG_SOFTLOCKUP_DETECTOR
952 {
953 .procname = "soft_watchdog",
954 .data = &soft_watchdog_user_enabled,
955 .maxlen = sizeof(int),
956 .mode = 0644,
957 .proc_handler = proc_soft_watchdog,
958 .extra1 = &zero,
959 .extra2 = &one,
960 },
961 {
962 .procname = "softlockup_panic",
963 .data = &softlockup_panic,
964 .maxlen = sizeof(int),
965 .mode = 0644,
966 .proc_handler = proc_dointvec_minmax,
967 .extra1 = &zero,
968 .extra2 = &one,
969 },
970 #ifdef CONFIG_SMP
971 {
972 .procname = "softlockup_all_cpu_backtrace",
973 .data = &sysctl_softlockup_all_cpu_backtrace,
974 .maxlen = sizeof(int),
975 .mode = 0644,
976 .proc_handler = proc_dointvec_minmax,
977 .extra1 = &zero,
978 .extra2 = &one,
979 },
980 #endif /* CONFIG_SMP */
981 #endif
982 #ifdef CONFIG_HARDLOCKUP_DETECTOR
983 {
984 .procname = "hardlockup_panic",
985 .data = &hardlockup_panic,
986 .maxlen = sizeof(int),
987 .mode = 0644,
988 .proc_handler = proc_dointvec_minmax,
989 .extra1 = &zero,
990 .extra2 = &one,
991 },
992 #ifdef CONFIG_SMP
993 {
994 .procname = "hardlockup_all_cpu_backtrace",
995 .data = &sysctl_hardlockup_all_cpu_backtrace,
996 .maxlen = sizeof(int),
997 .mode = 0644,
998 .proc_handler = proc_dointvec_minmax,
999 .extra1 = &zero,
1000 .extra2 = &one,
1001 },
1002 #endif /* CONFIG_SMP */
1003 #endif
1004 #endif
1005
1006 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
1007 {
1008 .procname = "unknown_nmi_panic",
1009 .data = &unknown_nmi_panic,
1010 .maxlen = sizeof (int),
1011 .mode = 0644,
1012 .proc_handler = proc_dointvec,
1013 },
1014 #endif
1015 #if defined(CONFIG_X86)
1016 {
1017 .procname = "panic_on_unrecovered_nmi",
1018 .data = &panic_on_unrecovered_nmi,
1019 .maxlen = sizeof(int),
1020 .mode = 0644,
1021 .proc_handler = proc_dointvec,
1022 },
1023 {
1024 .procname = "panic_on_io_nmi",
1025 .data = &panic_on_io_nmi,
1026 .maxlen = sizeof(int),
1027 .mode = 0644,
1028 .proc_handler = proc_dointvec,
1029 },
1030 #ifdef CONFIG_DEBUG_STACKOVERFLOW
1031 {
1032 .procname = "panic_on_stackoverflow",
1033 .data = &sysctl_panic_on_stackoverflow,
1034 .maxlen = sizeof(int),
1035 .mode = 0644,
1036 .proc_handler = proc_dointvec,
1037 },
1038 #endif
1039 {
1040 .procname = "bootloader_type",
1041 .data = &bootloader_type,
1042 .maxlen = sizeof (int),
1043 .mode = 0444,
1044 .proc_handler = proc_dointvec,
1045 },
1046 {
1047 .procname = "bootloader_version",
1048 .data = &bootloader_version,
1049 .maxlen = sizeof (int),
1050 .mode = 0444,
1051 .proc_handler = proc_dointvec,
1052 },
1053 {
1054 .procname = "io_delay_type",
1055 .data = &io_delay_type,
1056 .maxlen = sizeof(int),
1057 .mode = 0644,
1058 .proc_handler = proc_dointvec,
1059 },
1060 #endif
1061 #if defined(CONFIG_MMU)
1062 {
1063 .procname = "randomize_va_space",
1064 .data = &randomize_va_space,
1065 .maxlen = sizeof(int),
1066 .mode = 0644,
1067 .proc_handler = proc_dointvec,
1068 },
1069 #endif
1070 #if defined(CONFIG_S390) && defined(CONFIG_SMP)
1071 {
1072 .procname = "spin_retry",
1073 .data = &spin_retry,
1074 .maxlen = sizeof (int),
1075 .mode = 0644,
1076 .proc_handler = proc_dointvec,
1077 },
1078 #endif
1079 #if defined(CONFIG_ACPI_SLEEP) && defined(CONFIG_X86)
1080 {
1081 .procname = "acpi_video_flags",
1082 .data = &acpi_realmode_flags,
1083 .maxlen = sizeof (unsigned long),
1084 .mode = 0644,
1085 .proc_handler = proc_doulongvec_minmax,
1086 },
1087 #endif
1088 #ifdef CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN
1089 {
1090 .procname = "ignore-unaligned-usertrap",
1091 .data = &no_unaligned_warning,
1092 .maxlen = sizeof (int),
1093 .mode = 0644,
1094 .proc_handler = proc_dointvec,
1095 },
1096 #endif
1097 #ifdef CONFIG_IA64
1098 {
1099 .procname = "unaligned-dump-stack",
1100 .data = &unaligned_dump_stack,
1101 .maxlen = sizeof (int),
1102 .mode = 0644,
1103 .proc_handler = proc_dointvec,
1104 },
1105 #endif
1106 #ifdef CONFIG_DETECT_HUNG_TASK
1107 {
1108 .procname = "hung_task_panic",
1109 .data = &sysctl_hung_task_panic,
1110 .maxlen = sizeof(int),
1111 .mode = 0644,
1112 .proc_handler = proc_dointvec_minmax,
1113 .extra1 = &zero,
1114 .extra2 = &one,
1115 },
1116 {
1117 .procname = "hung_task_check_count",
1118 .data = &sysctl_hung_task_check_count,
1119 .maxlen = sizeof(int),
1120 .mode = 0644,
1121 .proc_handler = proc_dointvec_minmax,
1122 .extra1 = &zero,
1123 },
1124 {
1125 .procname = "hung_task_timeout_secs",
1126 .data = &sysctl_hung_task_timeout_secs,
1127 .maxlen = sizeof(unsigned long),
1128 .mode = 0644,
1129 .proc_handler = proc_dohung_task_timeout_secs,
1130 .extra2 = &hung_task_timeout_max,
1131 },
1132 {
1133 .procname = "hung_task_warnings",
1134 .data = &sysctl_hung_task_warnings,
1135 .maxlen = sizeof(int),
1136 .mode = 0644,
1137 .proc_handler = proc_dointvec_minmax,
1138 .extra1 = &neg_one,
1139 },
1140 #endif
1141 #ifdef CONFIG_RT_MUTEXES
1142 {
1143 .procname = "max_lock_depth",
1144 .data = &max_lock_depth,
1145 .maxlen = sizeof(int),
1146 .mode = 0644,
1147 .proc_handler = proc_dointvec,
1148 },
1149 #endif
1150 {
1151 .procname = "poweroff_cmd",
1152 .data = &poweroff_cmd,
1153 .maxlen = POWEROFF_CMD_PATH_LEN,
1154 .mode = 0644,
1155 .proc_handler = proc_dostring,
1156 },
1157 #ifdef CONFIG_KEYS
1158 {
1159 .procname = "keys",
1160 .mode = 0555,
1161 .child = key_sysctls,
1162 },
1163 #endif
1164 #ifdef CONFIG_PERF_EVENTS
1165 /*
1166 * User-space scripts rely on the existence of this file
1167 * as a feature check for perf_events being enabled.
1168 *
1169 * So it's an ABI, do not remove!
1170 */
1171 {
1172 .procname = "perf_event_paranoid",
1173 .data = &sysctl_perf_event_paranoid,
1174 .maxlen = sizeof(sysctl_perf_event_paranoid),
1175 .mode = 0644,
1176 .proc_handler = proc_dointvec,
1177 },
1178 {
1179 .procname = "perf_event_mlock_kb",
1180 .data = &sysctl_perf_event_mlock,
1181 .maxlen = sizeof(sysctl_perf_event_mlock),
1182 .mode = 0644,
1183 .proc_handler = proc_dointvec,
1184 },
1185 {
1186 .procname = "perf_event_max_sample_rate",
1187 .data = &sysctl_perf_event_sample_rate,
1188 .maxlen = sizeof(sysctl_perf_event_sample_rate),
1189 .mode = 0644,
1190 .proc_handler = perf_proc_update_handler,
1191 .extra1 = &one,
1192 },
1193 {
1194 .procname = "perf_cpu_time_max_percent",
1195 .data = &sysctl_perf_cpu_time_max_percent,
1196 .maxlen = sizeof(sysctl_perf_cpu_time_max_percent),
1197 .mode = 0644,
1198 .proc_handler = perf_cpu_time_max_percent_handler,
1199 .extra1 = &zero,
1200 .extra2 = &one_hundred,
1201 },
1202 {
1203 .procname = "perf_event_max_stack",
1204 .data = &sysctl_perf_event_max_stack,
1205 .maxlen = sizeof(sysctl_perf_event_max_stack),
1206 .mode = 0644,
1207 .proc_handler = perf_event_max_stack_handler,
1208 .extra1 = &zero,
1209 .extra2 = &six_hundred_forty_kb,
1210 },
1211 {
1212 .procname = "perf_event_max_contexts_per_stack",
1213 .data = &sysctl_perf_event_max_contexts_per_stack,
1214 .maxlen = sizeof(sysctl_perf_event_max_contexts_per_stack),
1215 .mode = 0644,
1216 .proc_handler = perf_event_max_stack_handler,
1217 .extra1 = &zero,
1218 .extra2 = &one_thousand,
1219 },
1220 #endif
1221 {
1222 .procname = "panic_on_warn",
1223 .data = &panic_on_warn,
1224 .maxlen = sizeof(int),
1225 .mode = 0644,
1226 .proc_handler = proc_dointvec_minmax,
1227 .extra1 = &zero,
1228 .extra2 = &one,
1229 },
1230 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
1231 {
1232 .procname = "timer_migration",
1233 .data = &sysctl_timer_migration,
1234 .maxlen = sizeof(unsigned int),
1235 .mode = 0644,
1236 .proc_handler = timer_migration_handler,
1237 .extra1 = &zero,
1238 .extra2 = &one,
1239 },
1240 #endif
1241 #ifdef CONFIG_BPF_SYSCALL
1242 {
1243 .procname = "unprivileged_bpf_disabled",
1244 .data = &sysctl_unprivileged_bpf_disabled,
1245 .maxlen = sizeof(sysctl_unprivileged_bpf_disabled),
1246 .mode = 0644,
1247 /* only handle a transition from default "0" to "1" */
1248 .proc_handler = proc_dointvec_minmax,
1249 .extra1 = &one,
1250 .extra2 = &one,
1251 },
1252 #endif
1253 #if defined(CONFIG_TREE_RCU) || defined(CONFIG_PREEMPT_RCU)
1254 {
1255 .procname = "panic_on_rcu_stall",
1256 .data = &sysctl_panic_on_rcu_stall,
1257 .maxlen = sizeof(sysctl_panic_on_rcu_stall),
1258 .mode = 0644,
1259 .proc_handler = proc_dointvec_minmax,
1260 .extra1 = &zero,
1261 .extra2 = &one,
1262 },
1263 #endif
1264 { }
1265 };
1266
1267 static struct ctl_table vm_table[] = {
1268 {
1269 .procname = "overcommit_memory",
1270 .data = &sysctl_overcommit_memory,
1271 .maxlen = sizeof(sysctl_overcommit_memory),
1272 .mode = 0644,
1273 .proc_handler = proc_dointvec_minmax,
1274 .extra1 = &zero,
1275 .extra2 = &two,
1276 },
1277 {
1278 .procname = "panic_on_oom",
1279 .data = &sysctl_panic_on_oom,
1280 .maxlen = sizeof(sysctl_panic_on_oom),
1281 .mode = 0644,
1282 .proc_handler = proc_dointvec_minmax,
1283 .extra1 = &zero,
1284 .extra2 = &two,
1285 },
1286 {
1287 .procname = "oom_kill_allocating_task",
1288 .data = &sysctl_oom_kill_allocating_task,
1289 .maxlen = sizeof(sysctl_oom_kill_allocating_task),
1290 .mode = 0644,
1291 .proc_handler = proc_dointvec,
1292 },
1293 {
1294 .procname = "oom_dump_tasks",
1295 .data = &sysctl_oom_dump_tasks,
1296 .maxlen = sizeof(sysctl_oom_dump_tasks),
1297 .mode = 0644,
1298 .proc_handler = proc_dointvec,
1299 },
1300 {
1301 .procname = "overcommit_ratio",
1302 .data = &sysctl_overcommit_ratio,
1303 .maxlen = sizeof(sysctl_overcommit_ratio),
1304 .mode = 0644,
1305 .proc_handler = overcommit_ratio_handler,
1306 },
1307 {
1308 .procname = "overcommit_kbytes",
1309 .data = &sysctl_overcommit_kbytes,
1310 .maxlen = sizeof(sysctl_overcommit_kbytes),
1311 .mode = 0644,
1312 .proc_handler = overcommit_kbytes_handler,
1313 },
1314 {
1315 .procname = "page-cluster",
1316 .data = &page_cluster,
1317 .maxlen = sizeof(int),
1318 .mode = 0644,
1319 .proc_handler = proc_dointvec_minmax,
1320 .extra1 = &zero,
1321 },
1322 {
1323 .procname = "dirty_background_ratio",
1324 .data = &dirty_background_ratio,
1325 .maxlen = sizeof(dirty_background_ratio),
1326 .mode = 0644,
1327 .proc_handler = dirty_background_ratio_handler,
1328 .extra1 = &zero,
1329 .extra2 = &one_hundred,
1330 },
1331 {
1332 .procname = "dirty_background_bytes",
1333 .data = &dirty_background_bytes,
1334 .maxlen = sizeof(dirty_background_bytes),
1335 .mode = 0644,
1336 .proc_handler = dirty_background_bytes_handler,
1337 .extra1 = &one_ul,
1338 },
1339 {
1340 .procname = "dirty_ratio",
1341 .data = &vm_dirty_ratio,
1342 .maxlen = sizeof(vm_dirty_ratio),
1343 .mode = 0644,
1344 .proc_handler = dirty_ratio_handler,
1345 .extra1 = &zero,
1346 .extra2 = &one_hundred,
1347 },
1348 {
1349 .procname = "dirty_bytes",
1350 .data = &vm_dirty_bytes,
1351 .maxlen = sizeof(vm_dirty_bytes),
1352 .mode = 0644,
1353 .proc_handler = dirty_bytes_handler,
1354 .extra1 = &dirty_bytes_min,
1355 },
1356 {
1357 .procname = "dirty_writeback_centisecs",
1358 .data = &dirty_writeback_interval,
1359 .maxlen = sizeof(dirty_writeback_interval),
1360 .mode = 0644,
1361 .proc_handler = dirty_writeback_centisecs_handler,
1362 },
1363 {
1364 .procname = "dirty_expire_centisecs",
1365 .data = &dirty_expire_interval,
1366 .maxlen = sizeof(dirty_expire_interval),
1367 .mode = 0644,
1368 .proc_handler = proc_dointvec_minmax,
1369 .extra1 = &zero,
1370 },
1371 {
1372 .procname = "dirtytime_expire_seconds",
1373 .data = &dirtytime_expire_interval,
1374 .maxlen = sizeof(dirty_expire_interval),
1375 .mode = 0644,
1376 .proc_handler = dirtytime_interval_handler,
1377 .extra1 = &zero,
1378 },
1379 {
1380 .procname = "nr_pdflush_threads",
1381 .mode = 0444 /* read-only */,
1382 .proc_handler = pdflush_proc_obsolete,
1383 },
1384 {
1385 .procname = "swappiness",
1386 .data = &vm_swappiness,
1387 .maxlen = sizeof(vm_swappiness),
1388 .mode = 0644,
1389 .proc_handler = proc_dointvec_minmax,
1390 .extra1 = &zero,
1391 .extra2 = &one_hundred,
1392 },
1393 #ifdef CONFIG_HUGETLB_PAGE
1394 {
1395 .procname = "nr_hugepages",
1396 .data = NULL,
1397 .maxlen = sizeof(unsigned long),
1398 .mode = 0644,
1399 .proc_handler = hugetlb_sysctl_handler,
1400 },
1401 #ifdef CONFIG_NUMA
1402 {
1403 .procname = "nr_hugepages_mempolicy",
1404 .data = NULL,
1405 .maxlen = sizeof(unsigned long),
1406 .mode = 0644,
1407 .proc_handler = &hugetlb_mempolicy_sysctl_handler,
1408 },
1409 #endif
1410 {
1411 .procname = "hugetlb_shm_group",
1412 .data = &sysctl_hugetlb_shm_group,
1413 .maxlen = sizeof(gid_t),
1414 .mode = 0644,
1415 .proc_handler = proc_dointvec,
1416 },
1417 {
1418 .procname = "hugepages_treat_as_movable",
1419 .data = &hugepages_treat_as_movable,
1420 .maxlen = sizeof(int),
1421 .mode = 0644,
1422 .proc_handler = proc_dointvec,
1423 },
1424 {
1425 .procname = "nr_overcommit_hugepages",
1426 .data = NULL,
1427 .maxlen = sizeof(unsigned long),
1428 .mode = 0644,
1429 .proc_handler = hugetlb_overcommit_handler,
1430 },
1431 #endif
1432 {
1433 .procname = "lowmem_reserve_ratio",
1434 .data = &sysctl_lowmem_reserve_ratio,
1435 .maxlen = sizeof(sysctl_lowmem_reserve_ratio),
1436 .mode = 0644,
1437 .proc_handler = lowmem_reserve_ratio_sysctl_handler,
1438 },
1439 {
1440 .procname = "drop_caches",
1441 .data = &sysctl_drop_caches,
1442 .maxlen = sizeof(int),
1443 .mode = 0644,
1444 .proc_handler = drop_caches_sysctl_handler,
1445 .extra1 = &one,
1446 .extra2 = &four,
1447 },
1448 #ifdef CONFIG_COMPACTION
1449 {
1450 .procname = "compact_memory",
1451 .data = &sysctl_compact_memory,
1452 .maxlen = sizeof(int),
1453 .mode = 0200,
1454 .proc_handler = sysctl_compaction_handler,
1455 },
1456 {
1457 .procname = "extfrag_threshold",
1458 .data = &sysctl_extfrag_threshold,
1459 .maxlen = sizeof(int),
1460 .mode = 0644,
1461 .proc_handler = sysctl_extfrag_handler,
1462 .extra1 = &min_extfrag_threshold,
1463 .extra2 = &max_extfrag_threshold,
1464 },
1465 {
1466 .procname = "compact_unevictable_allowed",
1467 .data = &sysctl_compact_unevictable_allowed,
1468 .maxlen = sizeof(int),
1469 .mode = 0644,
1470 .proc_handler = proc_dointvec,
1471 .extra1 = &zero,
1472 .extra2 = &one,
1473 },
1474
1475 #endif /* CONFIG_COMPACTION */
1476 {
1477 .procname = "min_free_kbytes",
1478 .data = &min_free_kbytes,
1479 .maxlen = sizeof(min_free_kbytes),
1480 .mode = 0644,
1481 .proc_handler = min_free_kbytes_sysctl_handler,
1482 .extra1 = &zero,
1483 },
1484 {
1485 .procname = "watermark_scale_factor",
1486 .data = &watermark_scale_factor,
1487 .maxlen = sizeof(watermark_scale_factor),
1488 .mode = 0644,
1489 .proc_handler = watermark_scale_factor_sysctl_handler,
1490 .extra1 = &one,
1491 .extra2 = &one_thousand,
1492 },
1493 {
1494 .procname = "extra_free_kbytes",
1495 .data = &extra_free_kbytes,
1496 .maxlen = sizeof(extra_free_kbytes),
1497 .mode = 0644,
1498 .proc_handler = min_free_kbytes_sysctl_handler,
1499 .extra1 = &zero,
1500 },
1501 {
1502 .procname = "percpu_pagelist_fraction",
1503 .data = &percpu_pagelist_fraction,
1504 .maxlen = sizeof(percpu_pagelist_fraction),
1505 .mode = 0644,
1506 .proc_handler = percpu_pagelist_fraction_sysctl_handler,
1507 .extra1 = &zero,
1508 },
1509 #ifdef CONFIG_MMU
1510 {
1511 .procname = "max_map_count",
1512 .data = &sysctl_max_map_count,
1513 .maxlen = sizeof(sysctl_max_map_count),
1514 .mode = 0644,
1515 .proc_handler = proc_dointvec_minmax,
1516 .extra1 = &zero,
1517 },
1518 #else
1519 {
1520 .procname = "nr_trim_pages",
1521 .data = &sysctl_nr_trim_pages,
1522 .maxlen = sizeof(sysctl_nr_trim_pages),
1523 .mode = 0644,
1524 .proc_handler = proc_dointvec_minmax,
1525 .extra1 = &zero,
1526 },
1527 #endif
1528 {
1529 .procname = "laptop_mode",
1530 .data = &laptop_mode,
1531 .maxlen = sizeof(laptop_mode),
1532 .mode = 0644,
1533 .proc_handler = proc_dointvec_jiffies,
1534 },
1535 {
1536 .procname = "block_dump",
1537 .data = &block_dump,
1538 .maxlen = sizeof(block_dump),
1539 .mode = 0644,
1540 .proc_handler = proc_dointvec,
1541 .extra1 = &zero,
1542 },
1543 {
1544 .procname = "vfs_cache_pressure",
1545 .data = &sysctl_vfs_cache_pressure,
1546 .maxlen = sizeof(sysctl_vfs_cache_pressure),
1547 .mode = 0644,
1548 .proc_handler = proc_dointvec,
1549 .extra1 = &zero,
1550 },
1551 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1552 {
1553 .procname = "legacy_va_layout",
1554 .data = &sysctl_legacy_va_layout,
1555 .maxlen = sizeof(sysctl_legacy_va_layout),
1556 .mode = 0644,
1557 .proc_handler = proc_dointvec,
1558 .extra1 = &zero,
1559 },
1560 #endif
1561 #ifdef CONFIG_NUMA
1562 {
1563 .procname = "zone_reclaim_mode",
1564 .data = &node_reclaim_mode,
1565 .maxlen = sizeof(node_reclaim_mode),
1566 .mode = 0644,
1567 .proc_handler = proc_dointvec,
1568 .extra1 = &zero,
1569 },
1570 {
1571 .procname = "min_unmapped_ratio",
1572 .data = &sysctl_min_unmapped_ratio,
1573 .maxlen = sizeof(sysctl_min_unmapped_ratio),
1574 .mode = 0644,
1575 .proc_handler = sysctl_min_unmapped_ratio_sysctl_handler,
1576 .extra1 = &zero,
1577 .extra2 = &one_hundred,
1578 },
1579 {
1580 .procname = "min_slab_ratio",
1581 .data = &sysctl_min_slab_ratio,
1582 .maxlen = sizeof(sysctl_min_slab_ratio),
1583 .mode = 0644,
1584 .proc_handler = sysctl_min_slab_ratio_sysctl_handler,
1585 .extra1 = &zero,
1586 .extra2 = &one_hundred,
1587 },
1588 #endif
1589 #ifdef CONFIG_SMP
1590 {
1591 .procname = "stat_interval",
1592 .data = &sysctl_stat_interval,
1593 .maxlen = sizeof(sysctl_stat_interval),
1594 .mode = 0644,
1595 .proc_handler = proc_dointvec_jiffies,
1596 },
1597 {
1598 .procname = "stat_refresh",
1599 .data = NULL,
1600 .maxlen = 0,
1601 .mode = 0600,
1602 .proc_handler = vmstat_refresh,
1603 },
1604 #endif
1605 #ifdef CONFIG_MMU
1606 {
1607 .procname = "mmap_min_addr",
1608 .data = &dac_mmap_min_addr,
1609 .maxlen = sizeof(unsigned long),
1610 .mode = 0644,
1611 .proc_handler = mmap_min_addr_handler,
1612 },
1613 #endif
1614 #ifdef CONFIG_NUMA
1615 {
1616 .procname = "numa_zonelist_order",
1617 .data = &numa_zonelist_order,
1618 .maxlen = NUMA_ZONELIST_ORDER_LEN,
1619 .mode = 0644,
1620 .proc_handler = numa_zonelist_order_handler,
1621 },
1622 #endif
1623 #if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \
1624 (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
1625 {
1626 .procname = "vdso_enabled",
1627 #ifdef CONFIG_X86_32
1628 .data = &vdso32_enabled,
1629 .maxlen = sizeof(vdso32_enabled),
1630 #else
1631 .data = &vdso_enabled,
1632 .maxlen = sizeof(vdso_enabled),
1633 #endif
1634 .mode = 0644,
1635 .proc_handler = proc_dointvec,
1636 .extra1 = &zero,
1637 },
1638 #endif
1639 #ifdef CONFIG_HIGHMEM
1640 {
1641 .procname = "highmem_is_dirtyable",
1642 .data = &vm_highmem_is_dirtyable,
1643 .maxlen = sizeof(vm_highmem_is_dirtyable),
1644 .mode = 0644,
1645 .proc_handler = proc_dointvec_minmax,
1646 .extra1 = &zero,
1647 .extra2 = &one,
1648 },
1649 #endif
1650 #ifdef CONFIG_MEMORY_FAILURE
1651 {
1652 .procname = "memory_failure_early_kill",
1653 .data = &sysctl_memory_failure_early_kill,
1654 .maxlen = sizeof(sysctl_memory_failure_early_kill),
1655 .mode = 0644,
1656 .proc_handler = proc_dointvec_minmax,
1657 .extra1 = &zero,
1658 .extra2 = &one,
1659 },
1660 {
1661 .procname = "memory_failure_recovery",
1662 .data = &sysctl_memory_failure_recovery,
1663 .maxlen = sizeof(sysctl_memory_failure_recovery),
1664 .mode = 0644,
1665 .proc_handler = proc_dointvec_minmax,
1666 .extra1 = &zero,
1667 .extra2 = &one,
1668 },
1669 #endif
1670 {
1671 .procname = "user_reserve_kbytes",
1672 .data = &sysctl_user_reserve_kbytes,
1673 .maxlen = sizeof(sysctl_user_reserve_kbytes),
1674 .mode = 0644,
1675 .proc_handler = proc_doulongvec_minmax,
1676 },
1677 {
1678 .procname = "admin_reserve_kbytes",
1679 .data = &sysctl_admin_reserve_kbytes,
1680 .maxlen = sizeof(sysctl_admin_reserve_kbytes),
1681 .mode = 0644,
1682 .proc_handler = proc_doulongvec_minmax,
1683 },
1684 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_BITS
1685 {
1686 .procname = "mmap_rnd_bits",
1687 .data = &mmap_rnd_bits,
1688 .maxlen = sizeof(mmap_rnd_bits),
1689 .mode = 0600,
1690 .proc_handler = proc_dointvec_minmax,
1691 .extra1 = (void *)&mmap_rnd_bits_min,
1692 .extra2 = (void *)&mmap_rnd_bits_max,
1693 },
1694 #endif
1695 #ifdef CONFIG_HAVE_ARCH_MMAP_RND_COMPAT_BITS
1696 {
1697 .procname = "mmap_rnd_compat_bits",
1698 .data = &mmap_rnd_compat_bits,
1699 .maxlen = sizeof(mmap_rnd_compat_bits),
1700 .mode = 0600,
1701 .proc_handler = proc_dointvec_minmax,
1702 .extra1 = (void *)&mmap_rnd_compat_bits_min,
1703 .extra2 = (void *)&mmap_rnd_compat_bits_max,
1704 },
1705 #endif
1706 { }
1707 };
1708
1709 static struct ctl_table fs_table[] = {
1710 {
1711 .procname = "inode-nr",
1712 .data = &inodes_stat,
1713 .maxlen = 2*sizeof(long),
1714 .mode = 0444,
1715 .proc_handler = proc_nr_inodes,
1716 },
1717 {
1718 .procname = "inode-state",
1719 .data = &inodes_stat,
1720 .maxlen = 7*sizeof(long),
1721 .mode = 0444,
1722 .proc_handler = proc_nr_inodes,
1723 },
1724 {
1725 .procname = "file-nr",
1726 .data = &files_stat,
1727 .maxlen = sizeof(files_stat),
1728 .mode = 0444,
1729 .proc_handler = proc_nr_files,
1730 },
1731 {
1732 .procname = "file-max",
1733 .data = &files_stat.max_files,
1734 .maxlen = sizeof(files_stat.max_files),
1735 .mode = 0644,
1736 .proc_handler = proc_doulongvec_minmax,
1737 },
1738 {
1739 .procname = "nr_open",
1740 .data = &sysctl_nr_open,
1741 .maxlen = sizeof(unsigned int),
1742 .mode = 0644,
1743 .proc_handler = proc_dointvec_minmax,
1744 .extra1 = &sysctl_nr_open_min,
1745 .extra2 = &sysctl_nr_open_max,
1746 },
1747 {
1748 .procname = "dentry-state",
1749 .data = &dentry_stat,
1750 .maxlen = 6*sizeof(long),
1751 .mode = 0444,
1752 .proc_handler = proc_nr_dentry,
1753 },
1754 {
1755 .procname = "overflowuid",
1756 .data = &fs_overflowuid,
1757 .maxlen = sizeof(int),
1758 .mode = 0644,
1759 .proc_handler = proc_dointvec_minmax,
1760 .extra1 = &minolduid,
1761 .extra2 = &maxolduid,
1762 },
1763 {
1764 .procname = "overflowgid",
1765 .data = &fs_overflowgid,
1766 .maxlen = sizeof(int),
1767 .mode = 0644,
1768 .proc_handler = proc_dointvec_minmax,
1769 .extra1 = &minolduid,
1770 .extra2 = &maxolduid,
1771 },
1772 #ifdef CONFIG_FILE_LOCKING
1773 {
1774 .procname = "leases-enable",
1775 .data = &leases_enable,
1776 .maxlen = sizeof(int),
1777 .mode = 0644,
1778 .proc_handler = proc_dointvec,
1779 },
1780 #endif
1781 #ifdef CONFIG_DNOTIFY
1782 {
1783 .procname = "dir-notify-enable",
1784 .data = &dir_notify_enable,
1785 .maxlen = sizeof(int),
1786 .mode = 0644,
1787 .proc_handler = proc_dointvec,
1788 },
1789 #endif
1790 #ifdef CONFIG_MMU
1791 #ifdef CONFIG_FILE_LOCKING
1792 {
1793 .procname = "lease-break-time",
1794 .data = &lease_break_time,
1795 .maxlen = sizeof(int),
1796 .mode = 0644,
1797 .proc_handler = proc_dointvec,
1798 },
1799 #endif
1800 #ifdef CONFIG_AIO
1801 {
1802 .procname = "aio-nr",
1803 .data = &aio_nr,
1804 .maxlen = sizeof(aio_nr),
1805 .mode = 0444,
1806 .proc_handler = proc_doulongvec_minmax,
1807 },
1808 {
1809 .procname = "aio-max-nr",
1810 .data = &aio_max_nr,
1811 .maxlen = sizeof(aio_max_nr),
1812 .mode = 0644,
1813 .proc_handler = proc_doulongvec_minmax,
1814 },
1815 #endif /* CONFIG_AIO */
1816 #ifdef CONFIG_INOTIFY_USER
1817 {
1818 .procname = "inotify",
1819 .mode = 0555,
1820 .child = inotify_table,
1821 },
1822 #endif
1823 #ifdef CONFIG_EPOLL
1824 {
1825 .procname = "epoll",
1826 .mode = 0555,
1827 .child = epoll_table,
1828 },
1829 #endif
1830 #endif
1831 {
1832 .procname = "protected_symlinks",
1833 .data = &sysctl_protected_symlinks,
1834 .maxlen = sizeof(int),
1835 .mode = 0600,
1836 .proc_handler = proc_dointvec_minmax,
1837 .extra1 = &zero,
1838 .extra2 = &one,
1839 },
1840 {
1841 .procname = "protected_hardlinks",
1842 .data = &sysctl_protected_hardlinks,
1843 .maxlen = sizeof(int),
1844 .mode = 0600,
1845 .proc_handler = proc_dointvec_minmax,
1846 .extra1 = &zero,
1847 .extra2 = &one,
1848 },
1849 {
1850 .procname = "protected_fifos",
1851 .data = &sysctl_protected_fifos,
1852 .maxlen = sizeof(int),
1853 .mode = 0600,
1854 .proc_handler = proc_dointvec_minmax,
1855 .extra1 = &zero,
1856 .extra2 = &two,
1857 },
1858 {
1859 .procname = "protected_regular",
1860 .data = &sysctl_protected_regular,
1861 .maxlen = sizeof(int),
1862 .mode = 0600,
1863 .proc_handler = proc_dointvec_minmax,
1864 .extra1 = &zero,
1865 .extra2 = &two,
1866 },
1867 {
1868 .procname = "suid_dumpable",
1869 .data = &suid_dumpable,
1870 .maxlen = sizeof(int),
1871 .mode = 0644,
1872 .proc_handler = proc_dointvec_minmax_coredump,
1873 .extra1 = &zero,
1874 .extra2 = &two,
1875 },
1876 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1877 {
1878 .procname = "binfmt_misc",
1879 .mode = 0555,
1880 .child = sysctl_mount_point,
1881 },
1882 #endif
1883 {
1884 .procname = "pipe-max-size",
1885 .data = &pipe_max_size,
1886 .maxlen = sizeof(pipe_max_size),
1887 .mode = 0644,
1888 .proc_handler = &pipe_proc_fn,
1889 .extra1 = &pipe_min_size,
1890 },
1891 {
1892 .procname = "pipe-user-pages-hard",
1893 .data = &pipe_user_pages_hard,
1894 .maxlen = sizeof(pipe_user_pages_hard),
1895 .mode = 0644,
1896 .proc_handler = proc_doulongvec_minmax,
1897 },
1898 {
1899 .procname = "pipe-user-pages-soft",
1900 .data = &pipe_user_pages_soft,
1901 .maxlen = sizeof(pipe_user_pages_soft),
1902 .mode = 0644,
1903 .proc_handler = proc_doulongvec_minmax,
1904 },
1905 {
1906 .procname = "mount-max",
1907 .data = &sysctl_mount_max,
1908 .maxlen = sizeof(unsigned int),
1909 .mode = 0644,
1910 .proc_handler = proc_dointvec_minmax,
1911 .extra1 = &one,
1912 },
1913 { }
1914 };
1915
1916 static struct ctl_table debug_table[] = {
1917 #ifdef CONFIG_SYSCTL_EXCEPTION_TRACE
1918 {
1919 .procname = "exception-trace",
1920 .data = &show_unhandled_signals,
1921 .maxlen = sizeof(int),
1922 .mode = 0644,
1923 .proc_handler = proc_dointvec
1924 },
1925 #endif
1926 #if defined(CONFIG_OPTPROBES)
1927 {
1928 .procname = "kprobes-optimization",
1929 .data = &sysctl_kprobes_optimization,
1930 .maxlen = sizeof(int),
1931 .mode = 0644,
1932 .proc_handler = proc_kprobes_optimization_handler,
1933 .extra1 = &zero,
1934 .extra2 = &one,
1935 },
1936 #endif
1937 { }
1938 };
1939
1940 static struct ctl_table dev_table[] = {
1941 { }
1942 };
1943
1944 int __init sysctl_init(void)
1945 {
1946 struct ctl_table_header *hdr;
1947
1948 hdr = register_sysctl_table(sysctl_base_table);
1949 kmemleak_not_leak(hdr);
1950 return 0;
1951 }
1952
1953 #endif /* CONFIG_SYSCTL */
1954
1955 /*
1956 * /proc/sys support
1957 */
1958
1959 #ifdef CONFIG_PROC_SYSCTL
1960
1961 static int _proc_do_string(char *data, int maxlen, int write,
1962 char __user *buffer,
1963 size_t *lenp, loff_t *ppos)
1964 {
1965 size_t len;
1966 char __user *p;
1967 char c;
1968
1969 if (!data || !maxlen || !*lenp) {
1970 *lenp = 0;
1971 return 0;
1972 }
1973
1974 if (write) {
1975 if (sysctl_writes_strict == SYSCTL_WRITES_STRICT) {
1976 /* Only continue writes not past the end of buffer. */
1977 len = strlen(data);
1978 if (len > maxlen - 1)
1979 len = maxlen - 1;
1980
1981 if (*ppos > len)
1982 return 0;
1983 len = *ppos;
1984 } else {
1985 /* Start writing from beginning of buffer. */
1986 len = 0;
1987 }
1988
1989 *ppos += *lenp;
1990 p = buffer;
1991 while ((p - buffer) < *lenp && len < maxlen - 1) {
1992 if (get_user(c, p++))
1993 return -EFAULT;
1994 if (c == 0 || c == '\n')
1995 break;
1996 data[len++] = c;
1997 }
1998 data[len] = 0;
1999 } else {
2000 len = strlen(data);
2001 if (len > maxlen)
2002 len = maxlen;
2003
2004 if (*ppos > len) {
2005 *lenp = 0;
2006 return 0;
2007 }
2008
2009 data += *ppos;
2010 len -= *ppos;
2011
2012 if (len > *lenp)
2013 len = *lenp;
2014 if (len)
2015 if (copy_to_user(buffer, data, len))
2016 return -EFAULT;
2017 if (len < *lenp) {
2018 if (put_user('\n', buffer + len))
2019 return -EFAULT;
2020 len++;
2021 }
2022 *lenp = len;
2023 *ppos += len;
2024 }
2025 return 0;
2026 }
2027
2028 static void warn_sysctl_write(struct ctl_table *table)
2029 {
2030 pr_warn_once("%s wrote to %s when file position was not 0!\n"
2031 "This will not be supported in the future. To silence this\n"
2032 "warning, set kernel.sysctl_writes_strict = -1\n",
2033 current->comm, table->procname);
2034 }
2035
2036 /**
2037 * proc_first_pos_non_zero_ignore - check if firs position is allowed
2038 * @ppos: file position
2039 * @table: the sysctl table
2040 *
2041 * Returns true if the first position is non-zero and the sysctl_writes_strict
2042 * mode indicates this is not allowed for numeric input types. String proc
2043 * hadlers can ignore the return value.
2044 */
2045 static bool proc_first_pos_non_zero_ignore(loff_t *ppos,
2046 struct ctl_table *table)
2047 {
2048 if (!*ppos)
2049 return false;
2050
2051 switch (sysctl_writes_strict) {
2052 case SYSCTL_WRITES_STRICT:
2053 return true;
2054 case SYSCTL_WRITES_WARN:
2055 warn_sysctl_write(table);
2056 return false;
2057 default:
2058 return false;
2059 }
2060 }
2061
2062 /**
2063 * proc_dostring - read a string sysctl
2064 * @table: the sysctl table
2065 * @write: %TRUE if this is a write to the sysctl file
2066 * @buffer: the user buffer
2067 * @lenp: the size of the user buffer
2068 * @ppos: file position
2069 *
2070 * Reads/writes a string from/to the user buffer. If the kernel
2071 * buffer provided is not large enough to hold the string, the
2072 * string is truncated. The copied string is %NULL-terminated.
2073 * If the string is being read by the user process, it is copied
2074 * and a newline '\n' is added. It is truncated if the buffer is
2075 * not large enough.
2076 *
2077 * Returns 0 on success.
2078 */
2079 int proc_dostring(struct ctl_table *table, int write,
2080 void __user *buffer, size_t *lenp, loff_t *ppos)
2081 {
2082 if (write)
2083 proc_first_pos_non_zero_ignore(ppos, table);
2084
2085 return _proc_do_string((char *)(table->data), table->maxlen, write,
2086 (char __user *)buffer, lenp, ppos);
2087 }
2088
2089 static size_t proc_skip_spaces(char **buf)
2090 {
2091 size_t ret;
2092 char *tmp = skip_spaces(*buf);
2093 ret = tmp - *buf;
2094 *buf = tmp;
2095 return ret;
2096 }
2097
2098 static void proc_skip_char(char **buf, size_t *size, const char v)
2099 {
2100 while (*size) {
2101 if (**buf != v)
2102 break;
2103 (*size)--;
2104 (*buf)++;
2105 }
2106 }
2107
2108 #define TMPBUFLEN 22
2109 /**
2110 * proc_get_long - reads an ASCII formatted integer from a user buffer
2111 *
2112 * @buf: a kernel buffer
2113 * @size: size of the kernel buffer
2114 * @val: this is where the number will be stored
2115 * @neg: set to %TRUE if number is negative
2116 * @perm_tr: a vector which contains the allowed trailers
2117 * @perm_tr_len: size of the perm_tr vector
2118 * @tr: pointer to store the trailer character
2119 *
2120 * In case of success %0 is returned and @buf and @size are updated with
2121 * the amount of bytes read. If @tr is non-NULL and a trailing
2122 * character exists (size is non-zero after returning from this
2123 * function), @tr is updated with the trailing character.
2124 */
2125 static int proc_get_long(char **buf, size_t *size,
2126 unsigned long *val, bool *neg,
2127 const char *perm_tr, unsigned perm_tr_len, char *tr)
2128 {
2129 int len;
2130 char *p, tmp[TMPBUFLEN];
2131
2132 if (!*size)
2133 return -EINVAL;
2134
2135 len = *size;
2136 if (len > TMPBUFLEN - 1)
2137 len = TMPBUFLEN - 1;
2138
2139 memcpy(tmp, *buf, len);
2140
2141 tmp[len] = 0;
2142 p = tmp;
2143 if (*p == '-' && *size > 1) {
2144 *neg = true;
2145 p++;
2146 } else
2147 *neg = false;
2148 if (!isdigit(*p))
2149 return -EINVAL;
2150
2151 *val = simple_strtoul(p, &p, 0);
2152
2153 len = p - tmp;
2154
2155 /* We don't know if the next char is whitespace thus we may accept
2156 * invalid integers (e.g. 1234...a) or two integers instead of one
2157 * (e.g. 123...1). So lets not allow such large numbers. */
2158 if (len == TMPBUFLEN - 1)
2159 return -EINVAL;
2160
2161 if (len < *size && perm_tr_len && !memchr(perm_tr, *p, perm_tr_len))
2162 return -EINVAL;
2163
2164 if (tr && (len < *size))
2165 *tr = *p;
2166
2167 *buf += len;
2168 *size -= len;
2169
2170 return 0;
2171 }
2172
2173 /**
2174 * proc_put_long - converts an integer to a decimal ASCII formatted string
2175 *
2176 * @buf: the user buffer
2177 * @size: the size of the user buffer
2178 * @val: the integer to be converted
2179 * @neg: sign of the number, %TRUE for negative
2180 *
2181 * In case of success %0 is returned and @buf and @size are updated with
2182 * the amount of bytes written.
2183 */
2184 static int proc_put_long(void __user **buf, size_t *size, unsigned long val,
2185 bool neg)
2186 {
2187 int len;
2188 char tmp[TMPBUFLEN], *p = tmp;
2189
2190 sprintf(p, "%s%lu", neg ? "-" : "", val);
2191 len = strlen(tmp);
2192 if (len > *size)
2193 len = *size;
2194 if (copy_to_user(*buf, tmp, len))
2195 return -EFAULT;
2196 *size -= len;
2197 *buf += len;
2198 return 0;
2199 }
2200 #undef TMPBUFLEN
2201
2202 static int proc_put_char(void __user **buf, size_t *size, char c)
2203 {
2204 if (*size) {
2205 char __user **buffer = (char __user **)buf;
2206 if (put_user(c, *buffer))
2207 return -EFAULT;
2208 (*size)--, (*buffer)++;
2209 *buf = *buffer;
2210 }
2211 return 0;
2212 }
2213
2214 static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
2215 int *valp,
2216 int write, void *data)
2217 {
2218 if (write) {
2219 if (*negp) {
2220 if (*lvalp > (unsigned long) INT_MAX + 1)
2221 return -EINVAL;
2222 *valp = -*lvalp;
2223 } else {
2224 if (*lvalp > (unsigned long) INT_MAX)
2225 return -EINVAL;
2226 *valp = *lvalp;
2227 }
2228 } else {
2229 int val = *valp;
2230 if (val < 0) {
2231 *negp = true;
2232 *lvalp = -(unsigned long)val;
2233 } else {
2234 *negp = false;
2235 *lvalp = (unsigned long)val;
2236 }
2237 }
2238 return 0;
2239 }
2240
2241 static int do_proc_douintvec_conv(unsigned long *lvalp,
2242 unsigned int *valp,
2243 int write, void *data)
2244 {
2245 if (write) {
2246 if (*lvalp > UINT_MAX)
2247 return -EINVAL;
2248 *valp = *lvalp;
2249 } else {
2250 unsigned int val = *valp;
2251 *lvalp = (unsigned long)val;
2252 }
2253 return 0;
2254 }
2255
2256 static const char proc_wspace_sep[] = { ' ', '\t', '\n' };
2257
2258 static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table,
2259 int write, void __user *buffer,
2260 size_t *lenp, loff_t *ppos,
2261 int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
2262 int write, void *data),
2263 void *data)
2264 {
2265 int *i, vleft, first = 1, err = 0;
2266 size_t left;
2267 char *kbuf = NULL, *p;
2268
2269 if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
2270 *lenp = 0;
2271 return 0;
2272 }
2273
2274 i = (int *) tbl_data;
2275 vleft = table->maxlen / sizeof(*i);
2276 left = *lenp;
2277
2278 if (!conv)
2279 conv = do_proc_dointvec_conv;
2280
2281 if (write) {
2282 if (proc_first_pos_non_zero_ignore(ppos, table))
2283 goto out;
2284
2285 if (left > PAGE_SIZE - 1)
2286 left = PAGE_SIZE - 1;
2287 p = kbuf = memdup_user_nul(buffer, left);
2288 if (IS_ERR(kbuf))
2289 return PTR_ERR(kbuf);
2290 }
2291
2292 for (; left && vleft--; i++, first=0) {
2293 unsigned long lval;
2294 bool neg;
2295
2296 if (write) {
2297 left -= proc_skip_spaces(&p);
2298
2299 if (!left)
2300 break;
2301 err = proc_get_long(&p, &left, &lval, &neg,
2302 proc_wspace_sep,
2303 sizeof(proc_wspace_sep), NULL);
2304 if (err)
2305 break;
2306 if (conv(&neg, &lval, i, 1, data)) {
2307 err = -EINVAL;
2308 break;
2309 }
2310 } else {
2311 if (conv(&neg, &lval, i, 0, data)) {
2312 err = -EINVAL;
2313 break;
2314 }
2315 if (!first)
2316 err = proc_put_char(&buffer, &left, '\t');
2317 if (err)
2318 break;
2319 err = proc_put_long(&buffer, &left, lval, neg);
2320 if (err)
2321 break;
2322 }
2323 }
2324
2325 if (!write && !first && left && !err)
2326 err = proc_put_char(&buffer, &left, '\n');
2327 if (write && !err && left)
2328 left -= proc_skip_spaces(&p);
2329 if (write) {
2330 kfree(kbuf);
2331 if (first)
2332 return err ? : -EINVAL;
2333 }
2334 *lenp -= left;
2335 out:
2336 *ppos += *lenp;
2337 return err;
2338 }
2339
2340 static int do_proc_dointvec(struct ctl_table *table, int write,
2341 void __user *buffer, size_t *lenp, loff_t *ppos,
2342 int (*conv)(bool *negp, unsigned long *lvalp, int *valp,
2343 int write, void *data),
2344 void *data)
2345 {
2346 return __do_proc_dointvec(table->data, table, write,
2347 buffer, lenp, ppos, conv, data);
2348 }
2349
2350 static int do_proc_douintvec_w(unsigned int *tbl_data,
2351 struct ctl_table *table,
2352 void __user *buffer,
2353 size_t *lenp, loff_t *ppos,
2354 int (*conv)(unsigned long *lvalp,
2355 unsigned int *valp,
2356 int write, void *data),
2357 void *data)
2358 {
2359 unsigned long lval;
2360 int err = 0;
2361 size_t left;
2362 bool neg;
2363 char *kbuf = NULL, *p;
2364
2365 left = *lenp;
2366
2367 if (proc_first_pos_non_zero_ignore(ppos, table))
2368 goto bail_early;
2369
2370 if (left > PAGE_SIZE - 1)
2371 left = PAGE_SIZE - 1;
2372
2373 p = kbuf = memdup_user_nul(buffer, left);
2374 if (IS_ERR(kbuf))
2375 return -EINVAL;
2376
2377 left -= proc_skip_spaces(&p);
2378 if (!left) {
2379 err = -EINVAL;
2380 goto out_free;
2381 }
2382
2383 err = proc_get_long(&p, &left, &lval, &neg,
2384 proc_wspace_sep,
2385 sizeof(proc_wspace_sep), NULL);
2386 if (err || neg) {
2387 err = -EINVAL;
2388 goto out_free;
2389 }
2390
2391 if (conv(&lval, tbl_data, 1, data)) {
2392 err = -EINVAL;
2393 goto out_free;
2394 }
2395
2396 if (!err && left)
2397 left -= proc_skip_spaces(&p);
2398
2399 out_free:
2400 kfree(kbuf);
2401 if (err)
2402 return -EINVAL;
2403
2404 return 0;
2405
2406 /* This is in keeping with old __do_proc_dointvec() */
2407 bail_early:
2408 *ppos += *lenp;
2409 return err;
2410 }
2411
2412 static int do_proc_douintvec_r(unsigned int *tbl_data, void __user *buffer,
2413 size_t *lenp, loff_t *ppos,
2414 int (*conv)(unsigned long *lvalp,
2415 unsigned int *valp,
2416 int write, void *data),
2417 void *data)
2418 {
2419 unsigned long lval;
2420 int err = 0;
2421 size_t left;
2422
2423 left = *lenp;
2424
2425 if (conv(&lval, tbl_data, 0, data)) {
2426 err = -EINVAL;
2427 goto out;
2428 }
2429
2430 err = proc_put_long(&buffer, &left, lval, false);
2431 if (err || !left)
2432 goto out;
2433
2434 err = proc_put_char(&buffer, &left, '\n');
2435
2436 out:
2437 *lenp -= left;
2438 *ppos += *lenp;
2439
2440 return err;
2441 }
2442
2443 static int __do_proc_douintvec(void *tbl_data, struct ctl_table *table,
2444 int write, void __user *buffer,
2445 size_t *lenp, loff_t *ppos,
2446 int (*conv)(unsigned long *lvalp,
2447 unsigned int *valp,
2448 int write, void *data),
2449 void *data)
2450 {
2451 unsigned int *i, vleft;
2452
2453 if (!tbl_data || !table->maxlen || !*lenp || (*ppos && !write)) {
2454 *lenp = 0;
2455 return 0;
2456 }
2457
2458 i = (unsigned int *) tbl_data;
2459 vleft = table->maxlen / sizeof(*i);
2460
2461 /*
2462 * Arrays are not supported, keep this simple. *Do not* add
2463 * support for them.
2464 */
2465 if (vleft != 1) {
2466 *lenp = 0;
2467 return -EINVAL;
2468 }
2469
2470 if (!conv)
2471 conv = do_proc_douintvec_conv;
2472
2473 if (write)
2474 return do_proc_douintvec_w(i, table, buffer, lenp, ppos,
2475 conv, data);
2476 return do_proc_douintvec_r(i, buffer, lenp, ppos, conv, data);
2477 }
2478
2479 static int do_proc_douintvec(struct ctl_table *table, int write,
2480 void __user *buffer, size_t *lenp, loff_t *ppos,
2481 int (*conv)(unsigned long *lvalp,
2482 unsigned int *valp,
2483 int write, void *data),
2484 void *data)
2485 {
2486 return __do_proc_douintvec(table->data, table, write,
2487 buffer, lenp, ppos, conv, data);
2488 }
2489
2490 /**
2491 * proc_dointvec - read a vector of integers
2492 * @table: the sysctl table
2493 * @write: %TRUE if this is a write to the sysctl file
2494 * @buffer: the user buffer
2495 * @lenp: the size of the user buffer
2496 * @ppos: file position
2497 *
2498 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2499 * values from/to the user buffer, treated as an ASCII string.
2500 *
2501 * Returns 0 on success.
2502 */
2503 int proc_dointvec(struct ctl_table *table, int write,
2504 void __user *buffer, size_t *lenp, loff_t *ppos)
2505 {
2506 return do_proc_dointvec(table, write, buffer, lenp, ppos, NULL, NULL);
2507 }
2508
2509 /**
2510 * proc_douintvec - read a vector of unsigned integers
2511 * @table: the sysctl table
2512 * @write: %TRUE if this is a write to the sysctl file
2513 * @buffer: the user buffer
2514 * @lenp: the size of the user buffer
2515 * @ppos: file position
2516 *
2517 * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
2518 * values from/to the user buffer, treated as an ASCII string.
2519 *
2520 * Returns 0 on success.
2521 */
2522 int proc_douintvec(struct ctl_table *table, int write,
2523 void __user *buffer, size_t *lenp, loff_t *ppos)
2524 {
2525 return do_proc_douintvec(table, write, buffer, lenp, ppos,
2526 do_proc_douintvec_conv, NULL);
2527 }
2528
2529 /*
2530 * Taint values can only be increased
2531 * This means we can safely use a temporary.
2532 */
2533 static int proc_taint(struct ctl_table *table, int write,
2534 void __user *buffer, size_t *lenp, loff_t *ppos)
2535 {
2536 struct ctl_table t;
2537 unsigned long tmptaint = get_taint();
2538 int err;
2539
2540 if (write && !capable(CAP_SYS_ADMIN))
2541 return -EPERM;
2542
2543 t = *table;
2544 t.data = &tmptaint;
2545 err = proc_doulongvec_minmax(&t, write, buffer, lenp, ppos);
2546 if (err < 0)
2547 return err;
2548
2549 if (write) {
2550 /*
2551 * Poor man's atomic or. Not worth adding a primitive
2552 * to everyone's atomic.h for this
2553 */
2554 int i;
2555 for (i = 0; i < BITS_PER_LONG && tmptaint >> i; i++) {
2556 if ((tmptaint >> i) & 1)
2557 add_taint(i, LOCKDEP_STILL_OK);
2558 }
2559 }
2560
2561 return err;
2562 }
2563
2564 #ifdef CONFIG_PRINTK
2565 static int proc_dointvec_minmax_sysadmin(struct ctl_table *table, int write,
2566 void __user *buffer, size_t *lenp, loff_t *ppos)
2567 {
2568 if (write && !capable(CAP_SYS_ADMIN))
2569 return -EPERM;
2570
2571 return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2572 }
2573 #endif
2574
2575 struct do_proc_dointvec_minmax_conv_param {
2576 int *min;
2577 int *max;
2578 };
2579
2580 static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
2581 int *valp,
2582 int write, void *data)
2583 {
2584 struct do_proc_dointvec_minmax_conv_param *param = data;
2585 if (write) {
2586 int val = *negp ? -*lvalp : *lvalp;
2587 if ((param->min && *param->min > val) ||
2588 (param->max && *param->max < val))
2589 return -EINVAL;
2590 *valp = val;
2591 } else {
2592 int val = *valp;
2593 if (val < 0) {
2594 *negp = true;
2595 *lvalp = -(unsigned long)val;
2596 } else {
2597 *negp = false;
2598 *lvalp = (unsigned long)val;
2599 }
2600 }
2601 return 0;
2602 }
2603
2604 /**
2605 * proc_dointvec_minmax - read a vector of integers with min/max values
2606 * @table: the sysctl table
2607 * @write: %TRUE if this is a write to the sysctl file
2608 * @buffer: the user buffer
2609 * @lenp: the size of the user buffer
2610 * @ppos: file position
2611 *
2612 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2613 * values from/to the user buffer, treated as an ASCII string.
2614 *
2615 * This routine will ensure the values are within the range specified by
2616 * table->extra1 (min) and table->extra2 (max).
2617 *
2618 * Returns 0 on success.
2619 */
2620 int proc_dointvec_minmax(struct ctl_table *table, int write,
2621 void __user *buffer, size_t *lenp, loff_t *ppos)
2622 {
2623 struct do_proc_dointvec_minmax_conv_param param = {
2624 .min = (int *) table->extra1,
2625 .max = (int *) table->extra2,
2626 };
2627 return do_proc_dointvec(table, write, buffer, lenp, ppos,
2628 do_proc_dointvec_minmax_conv, &param);
2629 }
2630
2631 struct do_proc_douintvec_minmax_conv_param {
2632 unsigned int *min;
2633 unsigned int *max;
2634 };
2635
2636 static int do_proc_douintvec_minmax_conv(unsigned long *lvalp,
2637 unsigned int *valp,
2638 int write, void *data)
2639 {
2640 struct do_proc_douintvec_minmax_conv_param *param = data;
2641
2642 if (write) {
2643 unsigned int val = *lvalp;
2644
2645 if ((param->min && *param->min > val) ||
2646 (param->max && *param->max < val))
2647 return -ERANGE;
2648
2649 if (*lvalp > UINT_MAX)
2650 return -EINVAL;
2651 *valp = val;
2652 } else {
2653 unsigned int val = *valp;
2654 *lvalp = (unsigned long) val;
2655 }
2656
2657 return 0;
2658 }
2659
2660 /**
2661 * proc_douintvec_minmax - read a vector of unsigned ints with min/max values
2662 * @table: the sysctl table
2663 * @write: %TRUE if this is a write to the sysctl file
2664 * @buffer: the user buffer
2665 * @lenp: the size of the user buffer
2666 * @ppos: file position
2667 *
2668 * Reads/writes up to table->maxlen/sizeof(unsigned int) unsigned integer
2669 * values from/to the user buffer, treated as an ASCII string. Negative
2670 * strings are not allowed.
2671 *
2672 * This routine will ensure the values are within the range specified by
2673 * table->extra1 (min) and table->extra2 (max). There is a final sanity
2674 * check for UINT_MAX to avoid having to support wrap around uses from
2675 * userspace.
2676 *
2677 * Returns 0 on success.
2678 */
2679 int proc_douintvec_minmax(struct ctl_table *table, int write,
2680 void __user *buffer, size_t *lenp, loff_t *ppos)
2681 {
2682 struct do_proc_douintvec_minmax_conv_param param = {
2683 .min = (unsigned int *) table->extra1,
2684 .max = (unsigned int *) table->extra2,
2685 };
2686 return do_proc_douintvec(table, write, buffer, lenp, ppos,
2687 do_proc_douintvec_minmax_conv, &param);
2688 }
2689
2690 static void validate_coredump_safety(void)
2691 {
2692 #ifdef CONFIG_COREDUMP
2693 if (suid_dumpable == SUID_DUMP_ROOT &&
2694 core_pattern[0] != '/' && core_pattern[0] != '|') {
2695 printk(KERN_WARNING
2696 "Unsafe core_pattern used with fs.suid_dumpable=2.\n"
2697 "Pipe handler or fully qualified core dump path required.\n"
2698 "Set kernel.core_pattern before fs.suid_dumpable.\n"
2699 );
2700 }
2701 #endif
2702 }
2703
2704 static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
2705 void __user *buffer, size_t *lenp, loff_t *ppos)
2706 {
2707 int error = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
2708 if (!error)
2709 validate_coredump_safety();
2710 return error;
2711 }
2712
2713 #ifdef CONFIG_COREDUMP
2714 static int proc_dostring_coredump(struct ctl_table *table, int write,
2715 void __user *buffer, size_t *lenp, loff_t *ppos)
2716 {
2717 int error = proc_dostring(table, write, buffer, lenp, ppos);
2718 if (!error)
2719 validate_coredump_safety();
2720 return error;
2721 }
2722 #endif
2723
2724 static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table, int write,
2725 void __user *buffer,
2726 size_t *lenp, loff_t *ppos,
2727 unsigned long convmul,
2728 unsigned long convdiv)
2729 {
2730 unsigned long *i, *min, *max;
2731 int vleft, first = 1, err = 0;
2732 size_t left;
2733 char *kbuf = NULL, *p;
2734
2735 if (!data || !table->maxlen || !*lenp || (*ppos && !write)) {
2736 *lenp = 0;
2737 return 0;
2738 }
2739
2740 i = (unsigned long *) data;
2741 min = (unsigned long *) table->extra1;
2742 max = (unsigned long *) table->extra2;
2743 vleft = table->maxlen / sizeof(unsigned long);
2744 left = *lenp;
2745
2746 if (write) {
2747 if (proc_first_pos_non_zero_ignore(ppos, table))
2748 goto out;
2749
2750 if (left > PAGE_SIZE - 1)
2751 left = PAGE_SIZE - 1;
2752 p = kbuf = memdup_user_nul(buffer, left);
2753 if (IS_ERR(kbuf))
2754 return PTR_ERR(kbuf);
2755 }
2756
2757 for (; left && vleft--; i++, first = 0) {
2758 unsigned long val;
2759
2760 if (write) {
2761 bool neg;
2762
2763 left -= proc_skip_spaces(&p);
2764
2765 err = proc_get_long(&p, &left, &val, &neg,
2766 proc_wspace_sep,
2767 sizeof(proc_wspace_sep), NULL);
2768 if (err)
2769 break;
2770 if (neg)
2771 continue;
2772 val = convmul * val / convdiv;
2773 if ((min && val < *min) || (max && val > *max))
2774 continue;
2775 *i = val;
2776 } else {
2777 val = convdiv * (*i) / convmul;
2778 if (!first) {
2779 err = proc_put_char(&buffer, &left, '\t');
2780 if (err)
2781 break;
2782 }
2783 err = proc_put_long(&buffer, &left, val, false);
2784 if (err)
2785 break;
2786 }
2787 }
2788
2789 if (!write && !first && left && !err)
2790 err = proc_put_char(&buffer, &left, '\n');
2791 if (write && !err)
2792 left -= proc_skip_spaces(&p);
2793 if (write) {
2794 kfree(kbuf);
2795 if (first)
2796 return err ? : -EINVAL;
2797 }
2798 *lenp -= left;
2799 out:
2800 *ppos += *lenp;
2801 return err;
2802 }
2803
2804 static int do_proc_doulongvec_minmax(struct ctl_table *table, int write,
2805 void __user *buffer,
2806 size_t *lenp, loff_t *ppos,
2807 unsigned long convmul,
2808 unsigned long convdiv)
2809 {
2810 return __do_proc_doulongvec_minmax(table->data, table, write,
2811 buffer, lenp, ppos, convmul, convdiv);
2812 }
2813
2814 /**
2815 * proc_doulongvec_minmax - read a vector of long integers with min/max values
2816 * @table: the sysctl table
2817 * @write: %TRUE if this is a write to the sysctl file
2818 * @buffer: the user buffer
2819 * @lenp: the size of the user buffer
2820 * @ppos: file position
2821 *
2822 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2823 * values from/to the user buffer, treated as an ASCII string.
2824 *
2825 * This routine will ensure the values are within the range specified by
2826 * table->extra1 (min) and table->extra2 (max).
2827 *
2828 * Returns 0 on success.
2829 */
2830 int proc_doulongvec_minmax(struct ctl_table *table, int write,
2831 void __user *buffer, size_t *lenp, loff_t *ppos)
2832 {
2833 return do_proc_doulongvec_minmax(table, write, buffer, lenp, ppos, 1l, 1l);
2834 }
2835
2836 /**
2837 * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
2838 * @table: the sysctl table
2839 * @write: %TRUE if this is a write to the sysctl file
2840 * @buffer: the user buffer
2841 * @lenp: the size of the user buffer
2842 * @ppos: file position
2843 *
2844 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2845 * values from/to the user buffer, treated as an ASCII string. The values
2846 * are treated as milliseconds, and converted to jiffies when they are stored.
2847 *
2848 * This routine will ensure the values are within the range specified by
2849 * table->extra1 (min) and table->extra2 (max).
2850 *
2851 * Returns 0 on success.
2852 */
2853 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
2854 void __user *buffer,
2855 size_t *lenp, loff_t *ppos)
2856 {
2857 return do_proc_doulongvec_minmax(table, write, buffer,
2858 lenp, ppos, HZ, 1000l);
2859 }
2860
2861
2862 static int do_proc_dointvec_jiffies_conv(bool *negp, unsigned long *lvalp,
2863 int *valp,
2864 int write, void *data)
2865 {
2866 if (write) {
2867 if (*lvalp > INT_MAX / HZ)
2868 return 1;
2869 *valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
2870 } else {
2871 int val = *valp;
2872 unsigned long lval;
2873 if (val < 0) {
2874 *negp = true;
2875 lval = -(unsigned long)val;
2876 } else {
2877 *negp = false;
2878 lval = (unsigned long)val;
2879 }
2880 *lvalp = lval / HZ;
2881 }
2882 return 0;
2883 }
2884
2885 static int do_proc_dointvec_userhz_jiffies_conv(bool *negp, unsigned long *lvalp,
2886 int *valp,
2887 int write, void *data)
2888 {
2889 if (write) {
2890 if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
2891 return 1;
2892 *valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
2893 } else {
2894 int val = *valp;
2895 unsigned long lval;
2896 if (val < 0) {
2897 *negp = true;
2898 lval = -(unsigned long)val;
2899 } else {
2900 *negp = false;
2901 lval = (unsigned long)val;
2902 }
2903 *lvalp = jiffies_to_clock_t(lval);
2904 }
2905 return 0;
2906 }
2907
2908 static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp,
2909 int *valp,
2910 int write, void *data)
2911 {
2912 if (write) {
2913 unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
2914
2915 if (jif > INT_MAX)
2916 return 1;
2917 *valp = (int)jif;
2918 } else {
2919 int val = *valp;
2920 unsigned long lval;
2921 if (val < 0) {
2922 *negp = true;
2923 lval = -(unsigned long)val;
2924 } else {
2925 *negp = false;
2926 lval = (unsigned long)val;
2927 }
2928 *lvalp = jiffies_to_msecs(lval);
2929 }
2930 return 0;
2931 }
2932
2933 /**
2934 * proc_dointvec_jiffies - read a vector of integers as seconds
2935 * @table: the sysctl table
2936 * @write: %TRUE if this is a write to the sysctl file
2937 * @buffer: the user buffer
2938 * @lenp: the size of the user buffer
2939 * @ppos: file position
2940 *
2941 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2942 * values from/to the user buffer, treated as an ASCII string.
2943 * The values read are assumed to be in seconds, and are converted into
2944 * jiffies.
2945 *
2946 * Returns 0 on success.
2947 */
2948 int proc_dointvec_jiffies(struct ctl_table *table, int write,
2949 void __user *buffer, size_t *lenp, loff_t *ppos)
2950 {
2951 return do_proc_dointvec(table,write,buffer,lenp,ppos,
2952 do_proc_dointvec_jiffies_conv,NULL);
2953 }
2954
2955 /**
2956 * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
2957 * @table: the sysctl table
2958 * @write: %TRUE if this is a write to the sysctl file
2959 * @buffer: the user buffer
2960 * @lenp: the size of the user buffer
2961 * @ppos: pointer to the file position
2962 *
2963 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2964 * values from/to the user buffer, treated as an ASCII string.
2965 * The values read are assumed to be in 1/USER_HZ seconds, and
2966 * are converted into jiffies.
2967 *
2968 * Returns 0 on success.
2969 */
2970 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
2971 void __user *buffer, size_t *lenp, loff_t *ppos)
2972 {
2973 return do_proc_dointvec(table,write,buffer,lenp,ppos,
2974 do_proc_dointvec_userhz_jiffies_conv,NULL);
2975 }
2976
2977 /**
2978 * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
2979 * @table: the sysctl table
2980 * @write: %TRUE if this is a write to the sysctl file
2981 * @buffer: the user buffer
2982 * @lenp: the size of the user buffer
2983 * @ppos: file position
2984 * @ppos: the current position in the file
2985 *
2986 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2987 * values from/to the user buffer, treated as an ASCII string.
2988 * The values read are assumed to be in 1/1000 seconds, and
2989 * are converted into jiffies.
2990 *
2991 * Returns 0 on success.
2992 */
2993 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
2994 void __user *buffer, size_t *lenp, loff_t *ppos)
2995 {
2996 return do_proc_dointvec(table, write, buffer, lenp, ppos,
2997 do_proc_dointvec_ms_jiffies_conv, NULL);
2998 }
2999
3000 static int proc_do_cad_pid(struct ctl_table *table, int write,
3001 void __user *buffer, size_t *lenp, loff_t *ppos)
3002 {
3003 struct pid *new_pid;
3004 pid_t tmp;
3005 int r;
3006
3007 tmp = pid_vnr(cad_pid);
3008
3009 r = __do_proc_dointvec(&tmp, table, write, buffer,
3010 lenp, ppos, NULL, NULL);
3011 if (r || !write)
3012 return r;
3013
3014 new_pid = find_get_pid(tmp);
3015 if (!new_pid)
3016 return -ESRCH;
3017
3018 put_pid(xchg(&cad_pid, new_pid));
3019 return 0;
3020 }
3021
3022 /**
3023 * proc_do_large_bitmap - read/write from/to a large bitmap
3024 * @table: the sysctl table
3025 * @write: %TRUE if this is a write to the sysctl file
3026 * @buffer: the user buffer
3027 * @lenp: the size of the user buffer
3028 * @ppos: file position
3029 *
3030 * The bitmap is stored at table->data and the bitmap length (in bits)
3031 * in table->maxlen.
3032 *
3033 * We use a range comma separated format (e.g. 1,3-4,10-10) so that
3034 * large bitmaps may be represented in a compact manner. Writing into
3035 * the file will clear the bitmap then update it with the given input.
3036 *
3037 * Returns 0 on success.
3038 */
3039 int proc_do_large_bitmap(struct ctl_table *table, int write,
3040 void __user *buffer, size_t *lenp, loff_t *ppos)
3041 {
3042 int err = 0;
3043 bool first = 1;
3044 size_t left = *lenp;
3045 unsigned long bitmap_len = table->maxlen;
3046 unsigned long *bitmap = *(unsigned long **) table->data;
3047 unsigned long *tmp_bitmap = NULL;
3048 char tr_a[] = { '-', ',', '\n' }, tr_b[] = { ',', '\n', 0 }, c;
3049
3050 if (!bitmap || !bitmap_len || !left || (*ppos && !write)) {
3051 *lenp = 0;
3052 return 0;
3053 }
3054
3055 if (write) {
3056 char *kbuf, *p;
3057
3058 if (left > PAGE_SIZE - 1)
3059 left = PAGE_SIZE - 1;
3060
3061 p = kbuf = memdup_user_nul(buffer, left);
3062 if (IS_ERR(kbuf))
3063 return PTR_ERR(kbuf);
3064
3065 tmp_bitmap = kzalloc(BITS_TO_LONGS(bitmap_len) * sizeof(unsigned long),
3066 GFP_KERNEL);
3067 if (!tmp_bitmap) {
3068 kfree(kbuf);
3069 return -ENOMEM;
3070 }
3071 proc_skip_char(&p, &left, '\n');
3072 while (!err && left) {
3073 unsigned long val_a, val_b;
3074 bool neg;
3075
3076 err = proc_get_long(&p, &left, &val_a, &neg, tr_a,
3077 sizeof(tr_a), &c);
3078 if (err)
3079 break;
3080 if (val_a >= bitmap_len || neg) {
3081 err = -EINVAL;
3082 break;
3083 }
3084
3085 val_b = val_a;
3086 if (left) {
3087 p++;
3088 left--;
3089 }
3090
3091 if (c == '-') {
3092 err = proc_get_long(&p, &left, &val_b,
3093 &neg, tr_b, sizeof(tr_b),
3094 &c);
3095 if (err)
3096 break;
3097 if (val_b >= bitmap_len || neg ||
3098 val_a > val_b) {
3099 err = -EINVAL;
3100 break;
3101 }
3102 if (left) {
3103 p++;
3104 left--;
3105 }
3106 }
3107
3108 bitmap_set(tmp_bitmap, val_a, val_b - val_a + 1);
3109 first = 0;
3110 proc_skip_char(&p, &left, '\n');
3111 }
3112 kfree(kbuf);
3113 } else {
3114 unsigned long bit_a, bit_b = 0;
3115
3116 while (left) {
3117 bit_a = find_next_bit(bitmap, bitmap_len, bit_b);
3118 if (bit_a >= bitmap_len)
3119 break;
3120 bit_b = find_next_zero_bit(bitmap, bitmap_len,
3121 bit_a + 1) - 1;
3122
3123 if (!first) {
3124 err = proc_put_char(&buffer, &left, ',');
3125 if (err)
3126 break;
3127 }
3128 err = proc_put_long(&buffer, &left, bit_a, false);
3129 if (err)
3130 break;
3131 if (bit_a != bit_b) {
3132 err = proc_put_char(&buffer, &left, '-');
3133 if (err)
3134 break;
3135 err = proc_put_long(&buffer, &left, bit_b, false);
3136 if (err)
3137 break;
3138 }
3139
3140 first = 0; bit_b++;
3141 }
3142 if (!err)
3143 err = proc_put_char(&buffer, &left, '\n');
3144 }
3145
3146 if (!err) {
3147 if (write) {
3148 if (*ppos)
3149 bitmap_or(bitmap, bitmap, tmp_bitmap, bitmap_len);
3150 else
3151 bitmap_copy(bitmap, tmp_bitmap, bitmap_len);
3152 }
3153 kfree(tmp_bitmap);
3154 *lenp -= left;
3155 *ppos += *lenp;
3156 return 0;
3157 } else {
3158 kfree(tmp_bitmap);
3159 return err;
3160 }
3161 }
3162
3163 #else /* CONFIG_PROC_SYSCTL */
3164
3165 int proc_dostring(struct ctl_table *table, int write,
3166 void __user *buffer, size_t *lenp, loff_t *ppos)
3167 {
3168 return -ENOSYS;
3169 }
3170
3171 int proc_dointvec(struct ctl_table *table, int write,
3172 void __user *buffer, size_t *lenp, loff_t *ppos)
3173 {
3174 return -ENOSYS;
3175 }
3176
3177 int proc_douintvec(struct ctl_table *table, int write,
3178 void __user *buffer, size_t *lenp, loff_t *ppos)
3179 {
3180 return -ENOSYS;
3181 }
3182
3183 int proc_dointvec_minmax(struct ctl_table *table, int write,
3184 void __user *buffer, size_t *lenp, loff_t *ppos)
3185 {
3186 return -ENOSYS;
3187 }
3188
3189 int proc_douintvec_minmax(struct ctl_table *table, int write,
3190 void __user *buffer, size_t *lenp, loff_t *ppos)
3191 {
3192 return -ENOSYS;
3193 }
3194
3195 int proc_dointvec_jiffies(struct ctl_table *table, int write,
3196 void __user *buffer, size_t *lenp, loff_t *ppos)
3197 {
3198 return -ENOSYS;
3199 }
3200
3201 int proc_dointvec_userhz_jiffies(struct ctl_table *table, int write,
3202 void __user *buffer, size_t *lenp, loff_t *ppos)
3203 {
3204 return -ENOSYS;
3205 }
3206
3207 int proc_dointvec_ms_jiffies(struct ctl_table *table, int write,
3208 void __user *buffer, size_t *lenp, loff_t *ppos)
3209 {
3210 return -ENOSYS;
3211 }
3212
3213 int proc_doulongvec_minmax(struct ctl_table *table, int write,
3214 void __user *buffer, size_t *lenp, loff_t *ppos)
3215 {
3216 return -ENOSYS;
3217 }
3218
3219 int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
3220 void __user *buffer,
3221 size_t *lenp, loff_t *ppos)
3222 {
3223 return -ENOSYS;
3224 }
3225
3226
3227 #endif /* CONFIG_PROC_SYSCTL */
3228
3229 /*
3230 * No sense putting this after each symbol definition, twice,
3231 * exception granted :-)
3232 */
3233 EXPORT_SYMBOL(proc_dointvec);
3234 EXPORT_SYMBOL(proc_douintvec);
3235 EXPORT_SYMBOL(proc_dointvec_jiffies);
3236 EXPORT_SYMBOL(proc_dointvec_minmax);
3237 EXPORT_SYMBOL_GPL(proc_douintvec_minmax);
3238 EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
3239 EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
3240 EXPORT_SYMBOL(proc_dostring);
3241 EXPORT_SYMBOL(proc_doulongvec_minmax);
3242 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);