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