projects / GitHub / mt8127 / android_kernel_alcatel_ttab.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
blame | history | raw | HEAD
audit: rework execve audit
[GitHub/mt8127/android_kernel_alcatel_ttab.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/mm.h>
23 #include <linux/swap.h>
24 #include <linux/slab.h>
25 #include <linux/sysctl.h>
26 #include <linux/proc_fs.h>
27 #include <linux/capability.h>
28 #include <linux/ctype.h>
29 #include <linux/utsname.h>
30 #include <linux/capability.h>
31 #include <linux/smp_lock.h>
32 #include <linux/fs.h>
33 #include <linux/init.h>
34 #include <linux/kernel.h>
35 #include <linux/kobject.h>
36 #include <linux/net.h>
37 #include <linux/sysrq.h>
38 #include <linux/highuid.h>
39 #include <linux/writeback.h>
40 #include <linux/hugetlb.h>
41 #include <linux/security.h>
42 #include <linux/initrd.h>
43 #include <linux/times.h>
44 #include <linux/limits.h>
45 #include <linux/dcache.h>
46 #include <linux/syscalls.h>
47 #include <linux/nfs_fs.h>
48 #include <linux/acpi.h>
49 #include <linux/reboot.h>
50
51 #include <asm/uaccess.h>
52 #include <asm/processor.h>
53
54 #ifdef CONFIG_X86
55 #include <asm/nmi.h>
56 #include <asm/stacktrace.h>
57 #endif
58
59 #if defined(CONFIG_SYSCTL)
60
61 /* External variables not in a header file. */
62 extern int C_A_D;
63 extern int print_fatal_signals;
64 extern int sysctl_overcommit_memory;
65 extern int sysctl_overcommit_ratio;
66 extern int sysctl_panic_on_oom;
67 extern int max_threads;
68 extern int core_uses_pid;
69 extern int suid_dumpable;
70 extern char core_pattern[];
71 extern int pid_max;
72 extern int min_free_kbytes;
73 extern int printk_ratelimit_jiffies;
74 extern int printk_ratelimit_burst;
75 extern int pid_max_min, pid_max_max;
76 extern int sysctl_drop_caches;
77 extern int percpu_pagelist_fraction;
78 extern int compat_log;
79 extern int maps_protect;
80 extern int sysctl_stat_interval;
81 extern int audit_argv_kb;
82
83 /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
84 static int maxolduid = 65535;
85 static int minolduid;
86 static int min_percpu_pagelist_fract = 8;
87
88 static int ngroups_max = NGROUPS_MAX;
89
90 #ifdef CONFIG_KMOD
91 extern char modprobe_path[];
92 #endif
93 #ifdef CONFIG_CHR_DEV_SG
94 extern int sg_big_buff;
95 #endif
96
97 #ifdef __sparc__
98 extern char reboot_command [];
99 extern int stop_a_enabled;
100 extern int scons_pwroff;
101 #endif
102
103 #ifdef __hppa__
104 extern int pwrsw_enabled;
105 extern int unaligned_enabled;
106 #endif
107
108 #ifdef CONFIG_S390
109 #ifdef CONFIG_MATHEMU
110 extern int sysctl_ieee_emulation_warnings;
111 #endif
112 extern int sysctl_userprocess_debug;
113 extern int spin_retry;
114 #endif
115
116 extern int sysctl_hz_timer;
117
118 #ifdef CONFIG_BSD_PROCESS_ACCT
119 extern int acct_parm[];
120 #endif
121
122 #ifdef CONFIG_IA64
123 extern int no_unaligned_warning;
124 #endif
125
126 #ifdef CONFIG_RT_MUTEXES
127 extern int max_lock_depth;
128 #endif
129
130 #ifdef CONFIG_SYSCTL_SYSCALL
131 static int parse_table(int __user *, int, void __user *, size_t __user *,
132 void __user *, size_t, ctl_table *);
133 #endif
134
135
136 #ifdef CONFIG_PROC_SYSCTL
137 static int proc_do_cad_pid(ctl_table *table, int write, struct file *filp,
138 void __user *buffer, size_t *lenp, loff_t *ppos);
139 static int proc_dointvec_taint(ctl_table *table, int write, struct file *filp,
140 void __user *buffer, size_t *lenp, loff_t *ppos);
141 #endif
142
143 static ctl_table root_table[];
144 static struct ctl_table_header root_table_header =
145 { root_table, LIST_HEAD_INIT(root_table_header.ctl_entry) };
146
147 static ctl_table kern_table[];
148 static ctl_table vm_table[];
149 static ctl_table fs_table[];
150 static ctl_table debug_table[];
151 static ctl_table dev_table[];
152 extern ctl_table random_table[];
153 #ifdef CONFIG_UNIX98_PTYS
154 extern ctl_table pty_table[];
155 #endif
156 #ifdef CONFIG_INOTIFY_USER
157 extern ctl_table inotify_table[];
158 #endif
159
160 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
161 int sysctl_legacy_va_layout;
162 #endif
163
164
165 /* The default sysctl tables: */
166
167 static ctl_table root_table[] = {
168 {
169 .ctl_name = CTL_KERN,
170 .procname = "kernel",
171 .mode = 0555,
172 .child = kern_table,
173 },
174 {
175 .ctl_name = CTL_VM,
176 .procname = "vm",
177 .mode = 0555,
178 .child = vm_table,
179 },
180 #ifdef CONFIG_NET
181 {
182 .ctl_name = CTL_NET,
183 .procname = "net",
184 .mode = 0555,
185 .child = net_table,
186 },
187 #endif
188 {
189 .ctl_name = CTL_FS,
190 .procname = "fs",
191 .mode = 0555,
192 .child = fs_table,
193 },
194 {
195 .ctl_name = CTL_DEBUG,
196 .procname = "debug",
197 .mode = 0555,
198 .child = debug_table,
199 },
200 {
201 .ctl_name = CTL_DEV,
202 .procname = "dev",
203 .mode = 0555,
204 .child = dev_table,
205 },
206 /*
207 * NOTE: do not add new entries to this table unless you have read
208 * Documentation/sysctl/ctl_unnumbered.txt
209 */
210 { .ctl_name = 0 }
211 };
212
213 #ifdef CONFIG_SCHED_DEBUG
214 static unsigned long min_sched_granularity_ns = 100000; /* 100 usecs */
215 static unsigned long max_sched_granularity_ns = 1000000000; /* 1 second */
216 static unsigned long min_wakeup_granularity_ns; /* 0 usecs */
217 static unsigned long max_wakeup_granularity_ns = 1000000000; /* 1 second */
218 #endif
219
220 static ctl_table kern_table[] = {
221 #ifdef CONFIG_SCHED_DEBUG
222 {
223 .ctl_name = CTL_UNNUMBERED,
224 .procname = "sched_granularity_ns",
225 .data = &sysctl_sched_granularity,
226 .maxlen = sizeof(unsigned int),
227 .mode = 0644,
228 .proc_handler = &proc_dointvec_minmax,
229 .strategy = &sysctl_intvec,
230 .extra1 = &min_sched_granularity_ns,
231 .extra2 = &max_sched_granularity_ns,
232 },
233 {
234 .ctl_name = CTL_UNNUMBERED,
235 .procname = "sched_wakeup_granularity_ns",
236 .data = &sysctl_sched_wakeup_granularity,
237 .maxlen = sizeof(unsigned int),
238 .mode = 0644,
239 .proc_handler = &proc_dointvec_minmax,
240 .strategy = &sysctl_intvec,
241 .extra1 = &min_wakeup_granularity_ns,
242 .extra2 = &max_wakeup_granularity_ns,
243 },
244 {
245 .ctl_name = CTL_UNNUMBERED,
246 .procname = "sched_batch_wakeup_granularity_ns",
247 .data = &sysctl_sched_batch_wakeup_granularity,
248 .maxlen = sizeof(unsigned int),
249 .mode = 0644,
250 .proc_handler = &proc_dointvec_minmax,
251 .strategy = &sysctl_intvec,
252 .extra1 = &min_wakeup_granularity_ns,
253 .extra2 = &max_wakeup_granularity_ns,
254 },
255 {
256 .ctl_name = CTL_UNNUMBERED,
257 .procname = "sched_stat_granularity_ns",
258 .data = &sysctl_sched_stat_granularity,
259 .maxlen = sizeof(unsigned int),
260 .mode = 0644,
261 .proc_handler = &proc_dointvec_minmax,
262 .strategy = &sysctl_intvec,
263 .extra1 = &min_wakeup_granularity_ns,
264 .extra2 = &max_wakeup_granularity_ns,
265 },
266 {
267 .ctl_name = CTL_UNNUMBERED,
268 .procname = "sched_runtime_limit_ns",
269 .data = &sysctl_sched_runtime_limit,
270 .maxlen = sizeof(unsigned int),
271 .mode = 0644,
272 .proc_handler = &proc_dointvec_minmax,
273 .strategy = &sysctl_intvec,
274 .extra1 = &min_sched_granularity_ns,
275 .extra2 = &max_sched_granularity_ns,
276 },
277 {
278 .ctl_name = CTL_UNNUMBERED,
279 .procname = "sched_child_runs_first",
280 .data = &sysctl_sched_child_runs_first,
281 .maxlen = sizeof(unsigned int),
282 .mode = 0644,
283 .proc_handler = &proc_dointvec,
284 },
285 {
286 .ctl_name = CTL_UNNUMBERED,
287 .procname = "sched_features",
288 .data = &sysctl_sched_features,
289 .maxlen = sizeof(unsigned int),
290 .mode = 0644,
291 .proc_handler = &proc_dointvec,
292 },
293 #endif
294 {
295 .ctl_name = KERN_PANIC,
296 .procname = "panic",
297 .data = &panic_timeout,
298 .maxlen = sizeof(int),
299 .mode = 0644,
300 .proc_handler = &proc_dointvec,
301 },
302 {
303 .ctl_name = KERN_CORE_USES_PID,
304 .procname = "core_uses_pid",
305 .data = &core_uses_pid,
306 .maxlen = sizeof(int),
307 .mode = 0644,
308 .proc_handler = &proc_dointvec,
309 },
310 #ifdef CONFIG_AUDITSYSCALL
311 {
312 .ctl_name = CTL_UNNUMBERED,
313 .procname = "audit_argv_kb",
314 .data = &audit_argv_kb,
315 .maxlen = sizeof(int),
316 .mode = 0644,
317 .proc_handler = &proc_dointvec,
318 },
319 #endif
320 {
321 .ctl_name = KERN_CORE_PATTERN,
322 .procname = "core_pattern",
323 .data = core_pattern,
324 .maxlen = CORENAME_MAX_SIZE,
325 .mode = 0644,
326 .proc_handler = &proc_dostring,
327 .strategy = &sysctl_string,
328 },
329 #ifdef CONFIG_PROC_SYSCTL
330 {
331 .ctl_name = KERN_TAINTED,
332 .procname = "tainted",
333 .data = &tainted,
334 .maxlen = sizeof(int),
335 .mode = 0644,
336 .proc_handler = &proc_dointvec_taint,
337 },
338 #endif
339 {
340 .ctl_name = KERN_CAP_BSET,
341 .procname = "cap-bound",
342 .data = &cap_bset,
343 .maxlen = sizeof(kernel_cap_t),
344 .mode = 0600,
345 .proc_handler = &proc_dointvec_bset,
346 },
347 #ifdef CONFIG_BLK_DEV_INITRD
348 {
349 .ctl_name = KERN_REALROOTDEV,
350 .procname = "real-root-dev",
351 .data = &real_root_dev,
352 .maxlen = sizeof(int),
353 .mode = 0644,
354 .proc_handler = &proc_dointvec,
355 },
356 #endif
357 {
358 .ctl_name = CTL_UNNUMBERED,
359 .procname = "print-fatal-signals",
360 .data = &print_fatal_signals,
361 .maxlen = sizeof(int),
362 .mode = 0644,
363 .proc_handler = &proc_dointvec,
364 },
365 #ifdef __sparc__
366 {
367 .ctl_name = KERN_SPARC_REBOOT,
368 .procname = "reboot-cmd",
369 .data = reboot_command,
370 .maxlen = 256,
371 .mode = 0644,
372 .proc_handler = &proc_dostring,
373 .strategy = &sysctl_string,
374 },
375 {
376 .ctl_name = KERN_SPARC_STOP_A,
377 .procname = "stop-a",
378 .data = &stop_a_enabled,
379 .maxlen = sizeof (int),
380 .mode = 0644,
381 .proc_handler = &proc_dointvec,
382 },
383 {
384 .ctl_name = KERN_SPARC_SCONS_PWROFF,
385 .procname = "scons-poweroff",
386 .data = &scons_pwroff,
387 .maxlen = sizeof (int),
388 .mode = 0644,
389 .proc_handler = &proc_dointvec,
390 },
391 #endif
392 #ifdef __hppa__
393 {
394 .ctl_name = KERN_HPPA_PWRSW,
395 .procname = "soft-power",
396 .data = &pwrsw_enabled,
397 .maxlen = sizeof (int),
398 .mode = 0644,
399 .proc_handler = &proc_dointvec,
400 },
401 {
402 .ctl_name = KERN_HPPA_UNALIGNED,
403 .procname = "unaligned-trap",
404 .data = &unaligned_enabled,
405 .maxlen = sizeof (int),
406 .mode = 0644,
407 .proc_handler = &proc_dointvec,
408 },
409 #endif
410 {
411 .ctl_name = KERN_CTLALTDEL,
412 .procname = "ctrl-alt-del",
413 .data = &C_A_D,
414 .maxlen = sizeof(int),
415 .mode = 0644,
416 .proc_handler = &proc_dointvec,
417 },
418 {
419 .ctl_name = KERN_PRINTK,
420 .procname = "printk",
421 .data = &console_loglevel,
422 .maxlen = 4*sizeof(int),
423 .mode = 0644,
424 .proc_handler = &proc_dointvec,
425 },
426 #ifdef CONFIG_KMOD
427 {
428 .ctl_name = KERN_MODPROBE,
429 .procname = "modprobe",
430 .data = &modprobe_path,
431 .maxlen = KMOD_PATH_LEN,
432 .mode = 0644,
433 .proc_handler = &proc_dostring,
434 .strategy = &sysctl_string,
435 },
436 #endif
437 #if defined(CONFIG_HOTPLUG) && defined(CONFIG_NET)
438 {
439 .ctl_name = KERN_HOTPLUG,
440 .procname = "hotplug",
441 .data = &uevent_helper,
442 .maxlen = UEVENT_HELPER_PATH_LEN,
443 .mode = 0644,
444 .proc_handler = &proc_dostring,
445 .strategy = &sysctl_string,
446 },
447 #endif
448 #ifdef CONFIG_CHR_DEV_SG
449 {
450 .ctl_name = KERN_SG_BIG_BUFF,
451 .procname = "sg-big-buff",
452 .data = &sg_big_buff,
453 .maxlen = sizeof (int),
454 .mode = 0444,
455 .proc_handler = &proc_dointvec,
456 },
457 #endif
458 #ifdef CONFIG_BSD_PROCESS_ACCT
459 {
460 .ctl_name = KERN_ACCT,
461 .procname = "acct",
462 .data = &acct_parm,
463 .maxlen = 3*sizeof(int),
464 .mode = 0644,
465 .proc_handler = &proc_dointvec,
466 },
467 #endif
468 #ifdef CONFIG_MAGIC_SYSRQ
469 {
470 .ctl_name = KERN_SYSRQ,
471 .procname = "sysrq",
472 .data = &__sysrq_enabled,
473 .maxlen = sizeof (int),
474 .mode = 0644,
475 .proc_handler = &proc_dointvec,
476 },
477 #endif
478 #ifdef CONFIG_PROC_SYSCTL
479 {
480 .ctl_name = KERN_CADPID,
481 .procname = "cad_pid",
482 .data = NULL,
483 .maxlen = sizeof (int),
484 .mode = 0600,
485 .proc_handler = &proc_do_cad_pid,
486 },
487 #endif
488 {
489 .ctl_name = KERN_MAX_THREADS,
490 .procname = "threads-max",
491 .data = &max_threads,
492 .maxlen = sizeof(int),
493 .mode = 0644,
494 .proc_handler = &proc_dointvec,
495 },
496 {
497 .ctl_name = KERN_RANDOM,
498 .procname = "random",
499 .mode = 0555,
500 .child = random_table,
501 },
502 #ifdef CONFIG_UNIX98_PTYS
503 {
504 .ctl_name = KERN_PTY,
505 .procname = "pty",
506 .mode = 0555,
507 .child = pty_table,
508 },
509 #endif
510 {
511 .ctl_name = KERN_OVERFLOWUID,
512 .procname = "overflowuid",
513 .data = &overflowuid,
514 .maxlen = sizeof(int),
515 .mode = 0644,
516 .proc_handler = &proc_dointvec_minmax,
517 .strategy = &sysctl_intvec,
518 .extra1 = &minolduid,
519 .extra2 = &maxolduid,
520 },
521 {
522 .ctl_name = KERN_OVERFLOWGID,
523 .procname = "overflowgid",
524 .data = &overflowgid,
525 .maxlen = sizeof(int),
526 .mode = 0644,
527 .proc_handler = &proc_dointvec_minmax,
528 .strategy = &sysctl_intvec,
529 .extra1 = &minolduid,
530 .extra2 = &maxolduid,
531 },
532 #ifdef CONFIG_S390
533 #ifdef CONFIG_MATHEMU
534 {
535 .ctl_name = KERN_IEEE_EMULATION_WARNINGS,
536 .procname = "ieee_emulation_warnings",
537 .data = &sysctl_ieee_emulation_warnings,
538 .maxlen = sizeof(int),
539 .mode = 0644,
540 .proc_handler = &proc_dointvec,
541 },
542 #endif
543 #ifdef CONFIG_NO_IDLE_HZ
544 {
545 .ctl_name = KERN_HZ_TIMER,
546 .procname = "hz_timer",
547 .data = &sysctl_hz_timer,
548 .maxlen = sizeof(int),
549 .mode = 0644,
550 .proc_handler = &proc_dointvec,
551 },
552 #endif
553 {
554 .ctl_name = KERN_S390_USER_DEBUG_LOGGING,
555 .procname = "userprocess_debug",
556 .data = &sysctl_userprocess_debug,
557 .maxlen = sizeof(int),
558 .mode = 0644,
559 .proc_handler = &proc_dointvec,
560 },
561 #endif
562 {
563 .ctl_name = KERN_PIDMAX,
564 .procname = "pid_max",
565 .data = &pid_max,
566 .maxlen = sizeof (int),
567 .mode = 0644,
568 .proc_handler = &proc_dointvec_minmax,
569 .strategy = sysctl_intvec,
570 .extra1 = &pid_max_min,
571 .extra2 = &pid_max_max,
572 },
573 {
574 .ctl_name = KERN_PANIC_ON_OOPS,
575 .procname = "panic_on_oops",
576 .data = &panic_on_oops,
577 .maxlen = sizeof(int),
578 .mode = 0644,
579 .proc_handler = &proc_dointvec,
580 },
581 {
582 .ctl_name = KERN_PRINTK_RATELIMIT,
583 .procname = "printk_ratelimit",
584 .data = &printk_ratelimit_jiffies,
585 .maxlen = sizeof(int),
586 .mode = 0644,
587 .proc_handler = &proc_dointvec_jiffies,
588 .strategy = &sysctl_jiffies,
589 },
590 {
591 .ctl_name = KERN_PRINTK_RATELIMIT_BURST,
592 .procname = "printk_ratelimit_burst",
593 .data = &printk_ratelimit_burst,
594 .maxlen = sizeof(int),
595 .mode = 0644,
596 .proc_handler = &proc_dointvec,
597 },
598 {
599 .ctl_name = KERN_NGROUPS_MAX,
600 .procname = "ngroups_max",
601 .data = &ngroups_max,
602 .maxlen = sizeof (int),
603 .mode = 0444,
604 .proc_handler = &proc_dointvec,
605 },
606 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
607 {
608 .ctl_name = KERN_UNKNOWN_NMI_PANIC,
609 .procname = "unknown_nmi_panic",
610 .data = &unknown_nmi_panic,
611 .maxlen = sizeof (int),
612 .mode = 0644,
613 .proc_handler = &proc_dointvec,
614 },
615 {
616 .ctl_name = KERN_NMI_WATCHDOG,
617 .procname = "nmi_watchdog",
618 .data = &nmi_watchdog_enabled,
619 .maxlen = sizeof (int),
620 .mode = 0644,
621 .proc_handler = &proc_nmi_enabled,
622 },
623 #endif
624 #if defined(CONFIG_X86)
625 {
626 .ctl_name = KERN_PANIC_ON_NMI,
627 .procname = "panic_on_unrecovered_nmi",
628 .data = &panic_on_unrecovered_nmi,
629 .maxlen = sizeof(int),
630 .mode = 0644,
631 .proc_handler = &proc_dointvec,
632 },
633 {
634 .ctl_name = KERN_BOOTLOADER_TYPE,
635 .procname = "bootloader_type",
636 .data = &bootloader_type,
637 .maxlen = sizeof (int),
638 .mode = 0444,
639 .proc_handler = &proc_dointvec,
640 },
641 {
642 .ctl_name = CTL_UNNUMBERED,
643 .procname = "kstack_depth_to_print",
644 .data = &kstack_depth_to_print,
645 .maxlen = sizeof(int),
646 .mode = 0644,
647 .proc_handler = &proc_dointvec,
648 },
649 #endif
650 #if defined(CONFIG_MMU)
651 {
652 .ctl_name = KERN_RANDOMIZE,
653 .procname = "randomize_va_space",
654 .data = &randomize_va_space,
655 .maxlen = sizeof(int),
656 .mode = 0644,
657 .proc_handler = &proc_dointvec,
658 },
659 #endif
660 #if defined(CONFIG_S390) && defined(CONFIG_SMP)
661 {
662 .ctl_name = KERN_SPIN_RETRY,
663 .procname = "spin_retry",
664 .data = &spin_retry,
665 .maxlen = sizeof (int),
666 .mode = 0644,
667 .proc_handler = &proc_dointvec,
668 },
669 #endif
670 #ifdef CONFIG_ACPI_SLEEP
671 {
672 .ctl_name = KERN_ACPI_VIDEO_FLAGS,
673 .procname = "acpi_video_flags",
674 .data = &acpi_realmode_flags,
675 .maxlen = sizeof (unsigned long),
676 .mode = 0644,
677 .proc_handler = &proc_doulongvec_minmax,
678 },
679 #endif
680 #ifdef CONFIG_IA64
681 {
682 .ctl_name = KERN_IA64_UNALIGNED,
683 .procname = "ignore-unaligned-usertrap",
684 .data = &no_unaligned_warning,
685 .maxlen = sizeof (int),
686 .mode = 0644,
687 .proc_handler = &proc_dointvec,
688 },
689 #endif
690 #ifdef CONFIG_COMPAT
691 {
692 .ctl_name = KERN_COMPAT_LOG,
693 .procname = "compat-log",
694 .data = &compat_log,
695 .maxlen = sizeof (int),
696 .mode = 0644,
697 .proc_handler = &proc_dointvec,
698 },
699 #endif
700 #ifdef CONFIG_RT_MUTEXES
701 {
702 .ctl_name = KERN_MAX_LOCK_DEPTH,
703 .procname = "max_lock_depth",
704 .data = &max_lock_depth,
705 .maxlen = sizeof(int),
706 .mode = 0644,
707 .proc_handler = &proc_dointvec,
708 },
709 #endif
710 #ifdef CONFIG_PROC_FS
711 {
712 .ctl_name = CTL_UNNUMBERED,
713 .procname = "maps_protect",
714 .data = &maps_protect,
715 .maxlen = sizeof(int),
716 .mode = 0644,
717 .proc_handler = &proc_dointvec,
718 },
719 #endif
720 {
721 .ctl_name = CTL_UNNUMBERED,
722 .procname = "poweroff_cmd",
723 .data = &poweroff_cmd,
724 .maxlen = POWEROFF_CMD_PATH_LEN,
725 .mode = 0644,
726 .proc_handler = &proc_dostring,
727 .strategy = &sysctl_string,
728 },
729
730 { .ctl_name = 0 }
731 };
732
733 /* Constants for minimum and maximum testing in vm_table.
734 We use these as one-element integer vectors. */
735 static int zero;
736 static int one_hundred = 100;
737
738
739 static ctl_table vm_table[] = {
740 {
741 .ctl_name = VM_OVERCOMMIT_MEMORY,
742 .procname = "overcommit_memory",
743 .data = &sysctl_overcommit_memory,
744 .maxlen = sizeof(sysctl_overcommit_memory),
745 .mode = 0644,
746 .proc_handler = &proc_dointvec,
747 },
748 {
749 .ctl_name = VM_PANIC_ON_OOM,
750 .procname = "panic_on_oom",
751 .data = &sysctl_panic_on_oom,
752 .maxlen = sizeof(sysctl_panic_on_oom),
753 .mode = 0644,
754 .proc_handler = &proc_dointvec,
755 },
756 {
757 .ctl_name = VM_OVERCOMMIT_RATIO,
758 .procname = "overcommit_ratio",
759 .data = &sysctl_overcommit_ratio,
760 .maxlen = sizeof(sysctl_overcommit_ratio),
761 .mode = 0644,
762 .proc_handler = &proc_dointvec,
763 },
764 {
765 .ctl_name = VM_PAGE_CLUSTER,
766 .procname = "page-cluster",
767 .data = &page_cluster,
768 .maxlen = sizeof(int),
769 .mode = 0644,
770 .proc_handler = &proc_dointvec,
771 },
772 {
773 .ctl_name = VM_DIRTY_BACKGROUND,
774 .procname = "dirty_background_ratio",
775 .data = &dirty_background_ratio,
776 .maxlen = sizeof(dirty_background_ratio),
777 .mode = 0644,
778 .proc_handler = &proc_dointvec_minmax,
779 .strategy = &sysctl_intvec,
780 .extra1 = &zero,
781 .extra2 = &one_hundred,
782 },
783 {
784 .ctl_name = VM_DIRTY_RATIO,
785 .procname = "dirty_ratio",
786 .data = &vm_dirty_ratio,
787 .maxlen = sizeof(vm_dirty_ratio),
788 .mode = 0644,
789 .proc_handler = &proc_dointvec_minmax,
790 .strategy = &sysctl_intvec,
791 .extra1 = &zero,
792 .extra2 = &one_hundred,
793 },
794 {
795 .ctl_name = VM_DIRTY_WB_CS,
796 .procname = "dirty_writeback_centisecs",
797 .data = &dirty_writeback_interval,
798 .maxlen = sizeof(dirty_writeback_interval),
799 .mode = 0644,
800 .proc_handler = &dirty_writeback_centisecs_handler,
801 },
802 {
803 .ctl_name = VM_DIRTY_EXPIRE_CS,
804 .procname = "dirty_expire_centisecs",
805 .data = &dirty_expire_interval,
806 .maxlen = sizeof(dirty_expire_interval),
807 .mode = 0644,
808 .proc_handler = &proc_dointvec_userhz_jiffies,
809 },
810 {
811 .ctl_name = VM_NR_PDFLUSH_THREADS,
812 .procname = "nr_pdflush_threads",
813 .data = &nr_pdflush_threads,
814 .maxlen = sizeof nr_pdflush_threads,
815 .mode = 0444 /* read-only*/,
816 .proc_handler = &proc_dointvec,
817 },
818 {
819 .ctl_name = VM_SWAPPINESS,
820 .procname = "swappiness",
821 .data = &vm_swappiness,
822 .maxlen = sizeof(vm_swappiness),
823 .mode = 0644,
824 .proc_handler = &proc_dointvec_minmax,
825 .strategy = &sysctl_intvec,
826 .extra1 = &zero,
827 .extra2 = &one_hundred,
828 },
829 #ifdef CONFIG_HUGETLB_PAGE
830 {
831 .ctl_name = VM_HUGETLB_PAGES,
832 .procname = "nr_hugepages",
833 .data = &max_huge_pages,
834 .maxlen = sizeof(unsigned long),
835 .mode = 0644,
836 .proc_handler = &hugetlb_sysctl_handler,
837 .extra1 = (void *)&hugetlb_zero,
838 .extra2 = (void *)&hugetlb_infinity,
839 },
840 {
841 .ctl_name = VM_HUGETLB_GROUP,
842 .procname = "hugetlb_shm_group",
843 .data = &sysctl_hugetlb_shm_group,
844 .maxlen = sizeof(gid_t),
845 .mode = 0644,
846 .proc_handler = &proc_dointvec,
847 },
848 {
849 .ctl_name = CTL_UNNUMBERED,
850 .procname = "hugepages_treat_as_movable",
851 .data = &hugepages_treat_as_movable,
852 .maxlen = sizeof(int),
853 .mode = 0644,
854 .proc_handler = &hugetlb_treat_movable_handler,
855 },
856 #endif
857 {
858 .ctl_name = VM_LOWMEM_RESERVE_RATIO,
859 .procname = "lowmem_reserve_ratio",
860 .data = &sysctl_lowmem_reserve_ratio,
861 .maxlen = sizeof(sysctl_lowmem_reserve_ratio),
862 .mode = 0644,
863 .proc_handler = &lowmem_reserve_ratio_sysctl_handler,
864 .strategy = &sysctl_intvec,
865 },
866 {
867 .ctl_name = VM_DROP_PAGECACHE,
868 .procname = "drop_caches",
869 .data = &sysctl_drop_caches,
870 .maxlen = sizeof(int),
871 .mode = 0644,
872 .proc_handler = drop_caches_sysctl_handler,
873 .strategy = &sysctl_intvec,
874 },
875 {
876 .ctl_name = VM_MIN_FREE_KBYTES,
877 .procname = "min_free_kbytes",
878 .data = &min_free_kbytes,
879 .maxlen = sizeof(min_free_kbytes),
880 .mode = 0644,
881 .proc_handler = &min_free_kbytes_sysctl_handler,
882 .strategy = &sysctl_intvec,
883 .extra1 = &zero,
884 },
885 {
886 .ctl_name = VM_PERCPU_PAGELIST_FRACTION,
887 .procname = "percpu_pagelist_fraction",
888 .data = &percpu_pagelist_fraction,
889 .maxlen = sizeof(percpu_pagelist_fraction),
890 .mode = 0644,
891 .proc_handler = &percpu_pagelist_fraction_sysctl_handler,
892 .strategy = &sysctl_intvec,
893 .extra1 = &min_percpu_pagelist_fract,
894 },
895 #ifdef CONFIG_MMU
896 {
897 .ctl_name = VM_MAX_MAP_COUNT,
898 .procname = "max_map_count",
899 .data = &sysctl_max_map_count,
900 .maxlen = sizeof(sysctl_max_map_count),
901 .mode = 0644,
902 .proc_handler = &proc_dointvec
903 },
904 #endif
905 {
906 .ctl_name = VM_LAPTOP_MODE,
907 .procname = "laptop_mode",
908 .data = &laptop_mode,
909 .maxlen = sizeof(laptop_mode),
910 .mode = 0644,
911 .proc_handler = &proc_dointvec_jiffies,
912 .strategy = &sysctl_jiffies,
913 },
914 {
915 .ctl_name = VM_BLOCK_DUMP,
916 .procname = "block_dump",
917 .data = &block_dump,
918 .maxlen = sizeof(block_dump),
919 .mode = 0644,
920 .proc_handler = &proc_dointvec,
921 .strategy = &sysctl_intvec,
922 .extra1 = &zero,
923 },
924 {
925 .ctl_name = VM_VFS_CACHE_PRESSURE,
926 .procname = "vfs_cache_pressure",
927 .data = &sysctl_vfs_cache_pressure,
928 .maxlen = sizeof(sysctl_vfs_cache_pressure),
929 .mode = 0644,
930 .proc_handler = &proc_dointvec,
931 .strategy = &sysctl_intvec,
932 .extra1 = &zero,
933 },
934 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
935 {
936 .ctl_name = VM_LEGACY_VA_LAYOUT,
937 .procname = "legacy_va_layout",
938 .data = &sysctl_legacy_va_layout,
939 .maxlen = sizeof(sysctl_legacy_va_layout),
940 .mode = 0644,
941 .proc_handler = &proc_dointvec,
942 .strategy = &sysctl_intvec,
943 .extra1 = &zero,
944 },
945 #endif
946 #ifdef CONFIG_NUMA
947 {
948 .ctl_name = VM_ZONE_RECLAIM_MODE,
949 .procname = "zone_reclaim_mode",
950 .data = &zone_reclaim_mode,
951 .maxlen = sizeof(zone_reclaim_mode),
952 .mode = 0644,
953 .proc_handler = &proc_dointvec,
954 .strategy = &sysctl_intvec,
955 .extra1 = &zero,
956 },
957 {
958 .ctl_name = VM_MIN_UNMAPPED,
959 .procname = "min_unmapped_ratio",
960 .data = &sysctl_min_unmapped_ratio,
961 .maxlen = sizeof(sysctl_min_unmapped_ratio),
962 .mode = 0644,
963 .proc_handler = &sysctl_min_unmapped_ratio_sysctl_handler,
964 .strategy = &sysctl_intvec,
965 .extra1 = &zero,
966 .extra2 = &one_hundred,
967 },
968 {
969 .ctl_name = VM_MIN_SLAB,
970 .procname = "min_slab_ratio",
971 .data = &sysctl_min_slab_ratio,
972 .maxlen = sizeof(sysctl_min_slab_ratio),
973 .mode = 0644,
974 .proc_handler = &sysctl_min_slab_ratio_sysctl_handler,
975 .strategy = &sysctl_intvec,
976 .extra1 = &zero,
977 .extra2 = &one_hundred,
978 },
979 #endif
980 #ifdef CONFIG_SMP
981 {
982 .ctl_name = CTL_UNNUMBERED,
983 .procname = "stat_interval",
984 .data = &sysctl_stat_interval,
985 .maxlen = sizeof(sysctl_stat_interval),
986 .mode = 0644,
987 .proc_handler = &proc_dointvec_jiffies,
988 .strategy = &sysctl_jiffies,
989 },
990 #endif
991 #ifdef CONFIG_SECURITY
992 {
993 .ctl_name = CTL_UNNUMBERED,
994 .procname = "mmap_min_addr",
995 .data = &mmap_min_addr,
996 .maxlen = sizeof(unsigned long),
997 .mode = 0644,
998 .proc_handler = &proc_doulongvec_minmax,
999 },
1000 #ifdef CONFIG_NUMA
1001 {
1002 .ctl_name = CTL_UNNUMBERED,
1003 .procname = "numa_zonelist_order",
1004 .data = &numa_zonelist_order,
1005 .maxlen = NUMA_ZONELIST_ORDER_LEN,
1006 .mode = 0644,
1007 .proc_handler = &numa_zonelist_order_handler,
1008 .strategy = &sysctl_string,
1009 },
1010 #endif
1011 #endif
1012 #if defined(CONFIG_X86_32) || \
1013 (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL))
1014 {
1015 .ctl_name = VM_VDSO_ENABLED,
1016 .procname = "vdso_enabled",
1017 .data = &vdso_enabled,
1018 .maxlen = sizeof(vdso_enabled),
1019 .mode = 0644,
1020 .proc_handler = &proc_dointvec,
1021 .strategy = &sysctl_intvec,
1022 .extra1 = &zero,
1023 },
1024 #endif
1025 /*
1026 * NOTE: do not add new entries to this table unless you have read
1027 * Documentation/sysctl/ctl_unnumbered.txt
1028 */
1029 { .ctl_name = 0 }
1030 };
1031
1032 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1033 static ctl_table binfmt_misc_table[] = {
1034 { .ctl_name = 0 }
1035 };
1036 #endif
1037
1038 static ctl_table fs_table[] = {
1039 {
1040 .ctl_name = FS_NRINODE,
1041 .procname = "inode-nr",
1042 .data = &inodes_stat,
1043 .maxlen = 2*sizeof(int),
1044 .mode = 0444,
1045 .proc_handler = &proc_dointvec,
1046 },
1047 {
1048 .ctl_name = FS_STATINODE,
1049 .procname = "inode-state",
1050 .data = &inodes_stat,
1051 .maxlen = 7*sizeof(int),
1052 .mode = 0444,
1053 .proc_handler = &proc_dointvec,
1054 },
1055 {
1056 .ctl_name = FS_NRFILE,
1057 .procname = "file-nr",
1058 .data = &files_stat,
1059 .maxlen = 3*sizeof(int),
1060 .mode = 0444,
1061 .proc_handler = &proc_nr_files,
1062 },
1063 {
1064 .ctl_name = FS_MAXFILE,
1065 .procname = "file-max",
1066 .data = &files_stat.max_files,
1067 .maxlen = sizeof(int),
1068 .mode = 0644,
1069 .proc_handler = &proc_dointvec,
1070 },
1071 {
1072 .ctl_name = FS_DENTRY,
1073 .procname = "dentry-state",
1074 .data = &dentry_stat,
1075 .maxlen = 6*sizeof(int),
1076 .mode = 0444,
1077 .proc_handler = &proc_dointvec,
1078 },
1079 {
1080 .ctl_name = FS_OVERFLOWUID,
1081 .procname = "overflowuid",
1082 .data = &fs_overflowuid,
1083 .maxlen = sizeof(int),
1084 .mode = 0644,
1085 .proc_handler = &proc_dointvec_minmax,
1086 .strategy = &sysctl_intvec,
1087 .extra1 = &minolduid,
1088 .extra2 = &maxolduid,
1089 },
1090 {
1091 .ctl_name = FS_OVERFLOWGID,
1092 .procname = "overflowgid",
1093 .data = &fs_overflowgid,
1094 .maxlen = sizeof(int),
1095 .mode = 0644,
1096 .proc_handler = &proc_dointvec_minmax,
1097 .strategy = &sysctl_intvec,
1098 .extra1 = &minolduid,
1099 .extra2 = &maxolduid,
1100 },
1101 {
1102 .ctl_name = FS_LEASES,
1103 .procname = "leases-enable",
1104 .data = &leases_enable,
1105 .maxlen = sizeof(int),
1106 .mode = 0644,
1107 .proc_handler = &proc_dointvec,
1108 },
1109 #ifdef CONFIG_DNOTIFY
1110 {
1111 .ctl_name = FS_DIR_NOTIFY,
1112 .procname = "dir-notify-enable",
1113 .data = &dir_notify_enable,
1114 .maxlen = sizeof(int),
1115 .mode = 0644,
1116 .proc_handler = &proc_dointvec,
1117 },
1118 #endif
1119 #ifdef CONFIG_MMU
1120 {
1121 .ctl_name = FS_LEASE_TIME,
1122 .procname = "lease-break-time",
1123 .data = &lease_break_time,
1124 .maxlen = sizeof(int),
1125 .mode = 0644,
1126 .proc_handler = &proc_dointvec,
1127 },
1128 {
1129 .ctl_name = FS_AIO_NR,
1130 .procname = "aio-nr",
1131 .data = &aio_nr,
1132 .maxlen = sizeof(aio_nr),
1133 .mode = 0444,
1134 .proc_handler = &proc_doulongvec_minmax,
1135 },
1136 {
1137 .ctl_name = FS_AIO_MAX_NR,
1138 .procname = "aio-max-nr",
1139 .data = &aio_max_nr,
1140 .maxlen = sizeof(aio_max_nr),
1141 .mode = 0644,
1142 .proc_handler = &proc_doulongvec_minmax,
1143 },
1144 #ifdef CONFIG_INOTIFY_USER
1145 {
1146 .ctl_name = FS_INOTIFY,
1147 .procname = "inotify",
1148 .mode = 0555,
1149 .child = inotify_table,
1150 },
1151 #endif
1152 #endif
1153 {
1154 .ctl_name = KERN_SETUID_DUMPABLE,
1155 .procname = "suid_dumpable",
1156 .data = &suid_dumpable,
1157 .maxlen = sizeof(int),
1158 .mode = 0644,
1159 .proc_handler = &proc_dointvec,
1160 },
1161 #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE)
1162 {
1163 .ctl_name = CTL_UNNUMBERED,
1164 .procname = "binfmt_misc",
1165 .mode = 0555,
1166 .child = binfmt_misc_table,
1167 },
1168 #endif
1169 /*
1170 * NOTE: do not add new entries to this table unless you have read
1171 * Documentation/sysctl/ctl_unnumbered.txt
1172 */
1173 { .ctl_name = 0 }
1174 };
1175
1176 static ctl_table debug_table[] = {
1177 { .ctl_name = 0 }
1178 };
1179
1180 static ctl_table dev_table[] = {
1181 { .ctl_name = 0 }
1182 };
1183
1184 static DEFINE_SPINLOCK(sysctl_lock);
1185
1186 /* called under sysctl_lock */
1187 static int use_table(struct ctl_table_header *p)
1188 {
1189 if (unlikely(p->unregistering))
1190 return 0;
1191 p->used++;
1192 return 1;
1193 }
1194
1195 /* called under sysctl_lock */
1196 static void unuse_table(struct ctl_table_header *p)
1197 {
1198 if (!--p->used)
1199 if (unlikely(p->unregistering))
1200 complete(p->unregistering);
1201 }
1202
1203 /* called under sysctl_lock, will reacquire if has to wait */
1204 static void start_unregistering(struct ctl_table_header *p)
1205 {
1206 /*
1207 * if p->used is 0, nobody will ever touch that entry again;
1208 * we'll eliminate all paths to it before dropping sysctl_lock
1209 */
1210 if (unlikely(p->used)) {
1211 struct completion wait;
1212 init_completion(&wait);
1213 p->unregistering = &wait;
1214 spin_unlock(&sysctl_lock);
1215 wait_for_completion(&wait);
1216 spin_lock(&sysctl_lock);
1217 }
1218 /*
1219 * do not remove from the list until nobody holds it; walking the
1220 * list in do_sysctl() relies on that.
1221 */
1222 list_del_init(&p->ctl_entry);
1223 }
1224
1225 void sysctl_head_finish(struct ctl_table_header *head)
1226 {
1227 if (!head)
1228 return;
1229 spin_lock(&sysctl_lock);
1230 unuse_table(head);
1231 spin_unlock(&sysctl_lock);
1232 }
1233
1234 struct ctl_table_header *sysctl_head_next(struct ctl_table_header *prev)
1235 {
1236 struct ctl_table_header *head;
1237 struct list_head *tmp;
1238 spin_lock(&sysctl_lock);
1239 if (prev) {
1240 tmp = &prev->ctl_entry;
1241 unuse_table(prev);
1242 goto next;
1243 }
1244 tmp = &root_table_header.ctl_entry;
1245 for (;;) {
1246 head = list_entry(tmp, struct ctl_table_header, ctl_entry);
1247
1248 if (!use_table(head))
1249 goto next;
1250 spin_unlock(&sysctl_lock);
1251 return head;
1252 next:
1253 tmp = tmp->next;
1254 if (tmp == &root_table_header.ctl_entry)
1255 break;
1256 }
1257 spin_unlock(&sysctl_lock);
1258 return NULL;
1259 }
1260
1261 #ifdef CONFIG_SYSCTL_SYSCALL
1262 int do_sysctl(int __user *name, int nlen, void __user *oldval, size_t __user *oldlenp,
1263 void __user *newval, size_t newlen)
1264 {
1265 struct ctl_table_header *head;
1266 int error = -ENOTDIR;
1267
1268 if (nlen <= 0 || nlen >= CTL_MAXNAME)
1269 return -ENOTDIR;
1270 if (oldval) {
1271 int old_len;
1272 if (!oldlenp || get_user(old_len, oldlenp))
1273 return -EFAULT;
1274 }
1275
1276 for (head = sysctl_head_next(NULL); head;
1277 head = sysctl_head_next(head)) {
1278 error = parse_table(name, nlen, oldval, oldlenp,
1279 newval, newlen, head->ctl_table);
1280 if (error != -ENOTDIR) {
1281 sysctl_head_finish(head);
1282 break;
1283 }
1284 }
1285 return error;
1286 }
1287
1288 asmlinkage long sys_sysctl(struct __sysctl_args __user *args)
1289 {
1290 struct __sysctl_args tmp;
1291 int error;
1292
1293 if (copy_from_user(&tmp, args, sizeof(tmp)))
1294 return -EFAULT;
1295
1296 lock_kernel();
1297 error = do_sysctl(tmp.name, tmp.nlen, tmp.oldval, tmp.oldlenp,
1298 tmp.newval, tmp.newlen);
1299 unlock_kernel();
1300 return error;
1301 }
1302 #endif /* CONFIG_SYSCTL_SYSCALL */
1303
1304 /*
1305 * sysctl_perm does NOT grant the superuser all rights automatically, because
1306 * some sysctl variables are readonly even to root.
1307 */
1308
1309 static int test_perm(int mode, int op)
1310 {
1311 if (!current->euid)
1312 mode >>= 6;
1313 else if (in_egroup_p(0))
1314 mode >>= 3;
1315 if ((mode & op & 0007) == op)
1316 return 0;
1317 return -EACCES;
1318 }
1319
1320 int sysctl_perm(ctl_table *table, int op)
1321 {
1322 int error;
1323 error = security_sysctl(table, op);
1324 if (error)
1325 return error;
1326 return test_perm(table->mode, op);
1327 }
1328
1329 #ifdef CONFIG_SYSCTL_SYSCALL
1330 static int parse_table(int __user *name, int nlen,
1331 void __user *oldval, size_t __user *oldlenp,
1332 void __user *newval, size_t newlen,
1333 ctl_table *table)
1334 {
1335 int n;
1336 repeat:
1337 if (!nlen)
1338 return -ENOTDIR;
1339 if (get_user(n, name))
1340 return -EFAULT;
1341 for ( ; table->ctl_name || table->procname; table++) {
1342 if (!table->ctl_name)
1343 continue;
1344 if (n == table->ctl_name) {
1345 int error;
1346 if (table->child) {
1347 if (sysctl_perm(table, 001))
1348 return -EPERM;
1349 name++;
1350 nlen--;
1351 table = table->child;
1352 goto repeat;
1353 }
1354 error = do_sysctl_strategy(table, name, nlen,
1355 oldval, oldlenp,
1356 newval, newlen);
1357 return error;
1358 }
1359 }
1360 return -ENOTDIR;
1361 }
1362
1363 /* Perform the actual read/write of a sysctl table entry. */
1364 int do_sysctl_strategy (ctl_table *table,
1365 int __user *name, int nlen,
1366 void __user *oldval, size_t __user *oldlenp,
1367 void __user *newval, size_t newlen)
1368 {
1369 int op = 0, rc;
1370 size_t len;
1371
1372 if (oldval)
1373 op |= 004;
1374 if (newval)
1375 op |= 002;
1376 if (sysctl_perm(table, op))
1377 return -EPERM;
1378
1379 if (table->strategy) {
1380 rc = table->strategy(table, name, nlen, oldval, oldlenp,
1381 newval, newlen);
1382 if (rc < 0)
1383 return rc;
1384 if (rc > 0)
1385 return 0;
1386 }
1387
1388 /* If there is no strategy routine, or if the strategy returns
1389 * zero, proceed with automatic r/w */
1390 if (table->data && table->maxlen) {
1391 if (oldval && oldlenp) {
1392 if (get_user(len, oldlenp))
1393 return -EFAULT;
1394 if (len) {
1395 if (len > table->maxlen)
1396 len = table->maxlen;
1397 if(copy_to_user(oldval, table->data, len))
1398 return -EFAULT;
1399 if(put_user(len, oldlenp))
1400 return -EFAULT;
1401 }
1402 }
1403 if (newval && newlen) {
1404 len = newlen;
1405 if (len > table->maxlen)
1406 len = table->maxlen;
1407 if(copy_from_user(table->data, newval, len))
1408 return -EFAULT;
1409 }
1410 }
1411 return 0;
1412 }
1413 #endif /* CONFIG_SYSCTL_SYSCALL */
1414
1415 static void sysctl_set_parent(struct ctl_table *parent, struct ctl_table *table)
1416 {
1417 for (; table->ctl_name || table->procname; table++) {
1418 table->parent = parent;
1419 if (table->child)
1420 sysctl_set_parent(table, table->child);
1421 }
1422 }
1423
1424 static __init int sysctl_init(void)
1425 {
1426 sysctl_set_parent(NULL, root_table);
1427 return 0;
1428 }
1429
1430 core_initcall(sysctl_init);
1431
1432 /**
1433 * register_sysctl_table - register a sysctl hierarchy
1434 * @table: the top-level table structure
1435 *
1436 * Register a sysctl table hierarchy. @table should be a filled in ctl_table
1437 * array. An entry with a ctl_name of 0 terminates the table.
1438 *
1439 * The members of the &ctl_table structure are used as follows:
1440 *
1441 * ctl_name - This is the numeric sysctl value used by sysctl(2). The number
1442 * must be unique within that level of sysctl
1443 *
1444 * procname - the name of the sysctl file under /proc/sys. Set to %NULL to not
1445 * enter a sysctl file
1446 *
1447 * data - a pointer to data for use by proc_handler
1448 *
1449 * maxlen - the maximum size in bytes of the data
1450 *
1451 * mode - the file permissions for the /proc/sys file, and for sysctl(2)
1452 *
1453 * child - a pointer to the child sysctl table if this entry is a directory, or
1454 * %NULL.
1455 *
1456 * proc_handler - the text handler routine (described below)
1457 *
1458 * strategy - the strategy routine (described below)
1459 *
1460 * de - for internal use by the sysctl routines
1461 *
1462 * extra1, extra2 - extra pointers usable by the proc handler routines
1463 *
1464 * Leaf nodes in the sysctl tree will be represented by a single file
1465 * under /proc; non-leaf nodes will be represented by directories.
1466 *
1467 * sysctl(2) can automatically manage read and write requests through
1468 * the sysctl table. The data and maxlen fields of the ctl_table
1469 * struct enable minimal validation of the values being written to be
1470 * performed, and the mode field allows minimal authentication.
1471 *
1472 * More sophisticated management can be enabled by the provision of a
1473 * strategy routine with the table entry. This will be called before
1474 * any automatic read or write of the data is performed.
1475 *
1476 * The strategy routine may return
1477 *
1478 * < 0 - Error occurred (error is passed to user process)
1479 *
1480 * 0 - OK - proceed with automatic read or write.
1481 *
1482 * > 0 - OK - read or write has been done by the strategy routine, so
1483 * return immediately.
1484 *
1485 * There must be a proc_handler routine for any terminal nodes
1486 * mirrored under /proc/sys (non-terminals are handled by a built-in
1487 * directory handler). Several default handlers are available to
1488 * cover common cases -
1489 *
1490 * proc_dostring(), proc_dointvec(), proc_dointvec_jiffies(),
1491 * proc_dointvec_userhz_jiffies(), proc_dointvec_minmax(),
1492 * proc_doulongvec_ms_jiffies_minmax(), proc_doulongvec_minmax()
1493 *
1494 * It is the handler's job to read the input buffer from user memory
1495 * and process it. The handler should return 0 on success.
1496 *
1497 * This routine returns %NULL on a failure to register, and a pointer
1498 * to the table header on success.
1499 */
1500 struct ctl_table_header *register_sysctl_table(ctl_table * table)
1501 {
1502 struct ctl_table_header *tmp;
1503 tmp = kmalloc(sizeof(struct ctl_table_header), GFP_KERNEL);
1504 if (!tmp)
1505 return NULL;
1506 tmp->ctl_table = table;
1507 INIT_LIST_HEAD(&tmp->ctl_entry);
1508 tmp->used = 0;
1509 tmp->unregistering = NULL;
1510 sysctl_set_parent(NULL, table);
1511 spin_lock(&sysctl_lock);
1512 list_add_tail(&tmp->ctl_entry, &root_table_header.ctl_entry);
1513 spin_unlock(&sysctl_lock);
1514 return tmp;
1515 }
1516
1517 /**
1518 * unregister_sysctl_table - unregister a sysctl table hierarchy
1519 * @header: the header returned from register_sysctl_table
1520 *
1521 * Unregisters the sysctl table and all children. proc entries may not
1522 * actually be removed until they are no longer used by anyone.
1523 */
1524 void unregister_sysctl_table(struct ctl_table_header * header)
1525 {
1526 might_sleep();
1527 spin_lock(&sysctl_lock);
1528 start_unregistering(header);
1529 spin_unlock(&sysctl_lock);
1530 kfree(header);
1531 }
1532
1533 #else /* !CONFIG_SYSCTL */
1534 struct ctl_table_header *register_sysctl_table(ctl_table * table)
1535 {
1536 return NULL;
1537 }
1538
1539 void unregister_sysctl_table(struct ctl_table_header * table)
1540 {
1541 }
1542
1543 #endif /* CONFIG_SYSCTL */
1544
1545 /*
1546 * /proc/sys support
1547 */
1548
1549 #ifdef CONFIG_PROC_SYSCTL
1550
1551 static int _proc_do_string(void* data, int maxlen, int write,
1552 struct file *filp, void __user *buffer,
1553 size_t *lenp, loff_t *ppos)
1554 {
1555 size_t len;
1556 char __user *p;
1557 char c;
1558
1559 if (!data || !maxlen || !*lenp) {
1560 *lenp = 0;
1561 return 0;
1562 }
1563
1564 if (write) {
1565 len = 0;
1566 p = buffer;
1567 while (len < *lenp) {
1568 if (get_user(c, p++))
1569 return -EFAULT;
1570 if (c == 0 || c == '\n')
1571 break;
1572 len++;
1573 }
1574 if (len >= maxlen)
1575 len = maxlen-1;
1576 if(copy_from_user(data, buffer, len))
1577 return -EFAULT;
1578 ((char *) data)[len] = 0;
1579 *ppos += *lenp;
1580 } else {
1581 len = strlen(data);
1582 if (len > maxlen)
1583 len = maxlen;
1584
1585 if (*ppos > len) {
1586 *lenp = 0;
1587 return 0;
1588 }
1589
1590 data += *ppos;
1591 len -= *ppos;
1592
1593 if (len > *lenp)
1594 len = *lenp;
1595 if (len)
1596 if(copy_to_user(buffer, data, len))
1597 return -EFAULT;
1598 if (len < *lenp) {
1599 if(put_user('\n', ((char __user *) buffer) + len))
1600 return -EFAULT;
1601 len++;
1602 }
1603 *lenp = len;
1604 *ppos += len;
1605 }
1606 return 0;
1607 }
1608
1609 /**
1610 * proc_dostring - read a string sysctl
1611 * @table: the sysctl table
1612 * @write: %TRUE if this is a write to the sysctl file
1613 * @filp: the file structure
1614 * @buffer: the user buffer
1615 * @lenp: the size of the user buffer
1616 * @ppos: file position
1617 *
1618 * Reads/writes a string from/to the user buffer. If the kernel
1619 * buffer provided is not large enough to hold the string, the
1620 * string is truncated. The copied string is %NULL-terminated.
1621 * If the string is being read by the user process, it is copied
1622 * and a newline '\n' is added. It is truncated if the buffer is
1623 * not large enough.
1624 *
1625 * Returns 0 on success.
1626 */
1627 int proc_dostring(ctl_table *table, int write, struct file *filp,
1628 void __user *buffer, size_t *lenp, loff_t *ppos)
1629 {
1630 return _proc_do_string(table->data, table->maxlen, write, filp,
1631 buffer, lenp, ppos);
1632 }
1633
1634
1635 static int do_proc_dointvec_conv(int *negp, unsigned long *lvalp,
1636 int *valp,
1637 int write, void *data)
1638 {
1639 if (write) {
1640 *valp = *negp ? -*lvalp : *lvalp;
1641 } else {
1642 int val = *valp;
1643 if (val < 0) {
1644 *negp = -1;
1645 *lvalp = (unsigned long)-val;
1646 } else {
1647 *negp = 0;
1648 *lvalp = (unsigned long)val;
1649 }
1650 }
1651 return 0;
1652 }
1653
1654 static int __do_proc_dointvec(void *tbl_data, ctl_table *table,
1655 int write, struct file *filp, void __user *buffer,
1656 size_t *lenp, loff_t *ppos,
1657 int (*conv)(int *negp, unsigned long *lvalp, int *valp,
1658 int write, void *data),
1659 void *data)
1660 {
1661 #define TMPBUFLEN 21
1662 int *i, vleft, first=1, neg, val;
1663 unsigned long lval;
1664 size_t left, len;
1665
1666 char buf[TMPBUFLEN], *p;
1667 char __user *s = buffer;
1668
1669 if (!tbl_data || !table->maxlen || !*lenp ||
1670 (*ppos && !write)) {
1671 *lenp = 0;
1672 return 0;
1673 }
1674
1675 i = (int *) tbl_data;
1676 vleft = table->maxlen / sizeof(*i);
1677 left = *lenp;
1678
1679 if (!conv)
1680 conv = do_proc_dointvec_conv;
1681
1682 for (; left && vleft--; i++, first=0) {
1683 if (write) {
1684 while (left) {
1685 char c;
1686 if (get_user(c, s))
1687 return -EFAULT;
1688 if (!isspace(c))
1689 break;
1690 left--;
1691 s++;
1692 }
1693 if (!left)
1694 break;
1695 neg = 0;
1696 len = left;
1697 if (len > sizeof(buf) - 1)
1698 len = sizeof(buf) - 1;
1699 if (copy_from_user(buf, s, len))
1700 return -EFAULT;
1701 buf[len] = 0;
1702 p = buf;
1703 if (*p == '-' && left > 1) {
1704 neg = 1;
1705 p++;
1706 }
1707 if (*p < '0' || *p > '9')
1708 break;
1709
1710 lval = simple_strtoul(p, &p, 0);
1711
1712 len = p-buf;
1713 if ((len < left) && *p && !isspace(*p))
1714 break;
1715 if (neg)
1716 val = -val;
1717 s += len;
1718 left -= len;
1719
1720 if (conv(&neg, &lval, i, 1, data))
1721 break;
1722 } else {
1723 p = buf;
1724 if (!first)
1725 *p++ = '\t';
1726
1727 if (conv(&neg, &lval, i, 0, data))
1728 break;
1729
1730 sprintf(p, "%s%lu", neg ? "-" : "", lval);
1731 len = strlen(buf);
1732 if (len > left)
1733 len = left;
1734 if(copy_to_user(s, buf, len))
1735 return -EFAULT;
1736 left -= len;
1737 s += len;
1738 }
1739 }
1740
1741 if (!write && !first && left) {
1742 if(put_user('\n', s))
1743 return -EFAULT;
1744 left--, s++;
1745 }
1746 if (write) {
1747 while (left) {
1748 char c;
1749 if (get_user(c, s++))
1750 return -EFAULT;
1751 if (!isspace(c))
1752 break;
1753 left--;
1754 }
1755 }
1756 if (write && first)
1757 return -EINVAL;
1758 *lenp -= left;
1759 *ppos += *lenp;
1760 return 0;
1761 #undef TMPBUFLEN
1762 }
1763
1764 static int do_proc_dointvec(ctl_table *table, int write, struct file *filp,
1765 void __user *buffer, size_t *lenp, loff_t *ppos,
1766 int (*conv)(int *negp, unsigned long *lvalp, int *valp,
1767 int write, void *data),
1768 void *data)
1769 {
1770 return __do_proc_dointvec(table->data, table, write, filp,
1771 buffer, lenp, ppos, conv, data);
1772 }
1773
1774 /**
1775 * proc_dointvec - read a vector of integers
1776 * @table: the sysctl table
1777 * @write: %TRUE if this is a write to the sysctl file
1778 * @filp: the file structure
1779 * @buffer: the user buffer
1780 * @lenp: the size of the user buffer
1781 * @ppos: file position
1782 *
1783 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1784 * values from/to the user buffer, treated as an ASCII string.
1785 *
1786 * Returns 0 on success.
1787 */
1788 int proc_dointvec(ctl_table *table, int write, struct file *filp,
1789 void __user *buffer, size_t *lenp, loff_t *ppos)
1790 {
1791 return do_proc_dointvec(table,write,filp,buffer,lenp,ppos,
1792 NULL,NULL);
1793 }
1794
1795 #define OP_SET 0
1796 #define OP_AND 1
1797 #define OP_OR 2
1798
1799 static int do_proc_dointvec_bset_conv(int *negp, unsigned long *lvalp,
1800 int *valp,
1801 int write, void *data)
1802 {
1803 int op = *(int *)data;
1804 if (write) {
1805 int val = *negp ? -*lvalp : *lvalp;
1806 switch(op) {
1807 case OP_SET: *valp = val; break;
1808 case OP_AND: *valp &= val; break;
1809 case OP_OR: *valp |= val; break;
1810 }
1811 } else {
1812 int val = *valp;
1813 if (val < 0) {
1814 *negp = -1;
1815 *lvalp = (unsigned long)-val;
1816 } else {
1817 *negp = 0;
1818 *lvalp = (unsigned long)val;
1819 }
1820 }
1821 return 0;
1822 }
1823
1824 /*
1825 * init may raise the set.
1826 */
1827
1828 int proc_dointvec_bset(ctl_table *table, int write, struct file *filp,
1829 void __user *buffer, size_t *lenp, loff_t *ppos)
1830 {
1831 int op;
1832
1833 if (write && !capable(CAP_SYS_MODULE)) {
1834 return -EPERM;
1835 }
1836
1837 op = is_init(current) ? OP_SET : OP_AND;
1838 return do_proc_dointvec(table,write,filp,buffer,lenp,ppos,
1839 do_proc_dointvec_bset_conv,&op);
1840 }
1841
1842 /*
1843 * Taint values can only be increased
1844 */
1845 static int proc_dointvec_taint(ctl_table *table, int write, struct file *filp,
1846 void __user *buffer, size_t *lenp, loff_t *ppos)
1847 {
1848 int op;
1849
1850 if (write && !capable(CAP_SYS_ADMIN))
1851 return -EPERM;
1852
1853 op = OP_OR;
1854 return do_proc_dointvec(table,write,filp,buffer,lenp,ppos,
1855 do_proc_dointvec_bset_conv,&op);
1856 }
1857
1858 struct do_proc_dointvec_minmax_conv_param {
1859 int *min;
1860 int *max;
1861 };
1862
1863 static int do_proc_dointvec_minmax_conv(int *negp, unsigned long *lvalp,
1864 int *valp,
1865 int write, void *data)
1866 {
1867 struct do_proc_dointvec_minmax_conv_param *param = data;
1868 if (write) {
1869 int val = *negp ? -*lvalp : *lvalp;
1870 if ((param->min && *param->min > val) ||
1871 (param->max && *param->max < val))
1872 return -EINVAL;
1873 *valp = val;
1874 } else {
1875 int val = *valp;
1876 if (val < 0) {
1877 *negp = -1;
1878 *lvalp = (unsigned long)-val;
1879 } else {
1880 *negp = 0;
1881 *lvalp = (unsigned long)val;
1882 }
1883 }
1884 return 0;
1885 }
1886
1887 /**
1888 * proc_dointvec_minmax - read a vector of integers with min/max values
1889 * @table: the sysctl table
1890 * @write: %TRUE if this is a write to the sysctl file
1891 * @filp: the file structure
1892 * @buffer: the user buffer
1893 * @lenp: the size of the user buffer
1894 * @ppos: file position
1895 *
1896 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
1897 * values from/to the user buffer, treated as an ASCII string.
1898 *
1899 * This routine will ensure the values are within the range specified by
1900 * table->extra1 (min) and table->extra2 (max).
1901 *
1902 * Returns 0 on success.
1903 */
1904 int proc_dointvec_minmax(ctl_table *table, int write, struct file *filp,
1905 void __user *buffer, size_t *lenp, loff_t *ppos)
1906 {
1907 struct do_proc_dointvec_minmax_conv_param param = {
1908 .min = (int *) table->extra1,
1909 .max = (int *) table->extra2,
1910 };
1911 return do_proc_dointvec(table, write, filp, buffer, lenp, ppos,
1912 do_proc_dointvec_minmax_conv, &param);
1913 }
1914
1915 static int __do_proc_doulongvec_minmax(void *data, ctl_table *table, int write,
1916 struct file *filp,
1917 void __user *buffer,
1918 size_t *lenp, loff_t *ppos,
1919 unsigned long convmul,
1920 unsigned long convdiv)
1921 {
1922 #define TMPBUFLEN 21
1923 unsigned long *i, *min, *max, val;
1924 int vleft, first=1, neg;
1925 size_t len, left;
1926 char buf[TMPBUFLEN], *p;
1927 char __user *s = buffer;
1928
1929 if (!data || !table->maxlen || !*lenp ||
1930 (*ppos && !write)) {
1931 *lenp = 0;
1932 return 0;
1933 }
1934
1935 i = (unsigned long *) data;
1936 min = (unsigned long *) table->extra1;
1937 max = (unsigned long *) table->extra2;
1938 vleft = table->maxlen / sizeof(unsigned long);
1939 left = *lenp;
1940
1941 for (; left && vleft--; i++, min++, max++, first=0) {
1942 if (write) {
1943 while (left) {
1944 char c;
1945 if (get_user(c, s))
1946 return -EFAULT;
1947 if (!isspace(c))
1948 break;
1949 left--;
1950 s++;
1951 }
1952 if (!left)
1953 break;
1954 neg = 0;
1955 len = left;
1956 if (len > TMPBUFLEN-1)
1957 len = TMPBUFLEN-1;
1958 if (copy_from_user(buf, s, len))
1959 return -EFAULT;
1960 buf[len] = 0;
1961 p = buf;
1962 if (*p == '-' && left > 1) {
1963 neg = 1;
1964 p++;
1965 }
1966 if (*p < '0' || *p > '9')
1967 break;
1968 val = simple_strtoul(p, &p, 0) * convmul / convdiv ;
1969 len = p-buf;
1970 if ((len < left) && *p && !isspace(*p))
1971 break;
1972 if (neg)
1973 val = -val;
1974 s += len;
1975 left -= len;
1976
1977 if(neg)
1978 continue;
1979 if ((min && val < *min) || (max && val > *max))
1980 continue;
1981 *i = val;
1982 } else {
1983 p = buf;
1984 if (!first)
1985 *p++ = '\t';
1986 sprintf(p, "%lu", convdiv * (*i) / convmul);
1987 len = strlen(buf);
1988 if (len > left)
1989 len = left;
1990 if(copy_to_user(s, buf, len))
1991 return -EFAULT;
1992 left -= len;
1993 s += len;
1994 }
1995 }
1996
1997 if (!write && !first && left) {
1998 if(put_user('\n', s))
1999 return -EFAULT;
2000 left--, s++;
2001 }
2002 if (write) {
2003 while (left) {
2004 char c;
2005 if (get_user(c, s++))
2006 return -EFAULT;
2007 if (!isspace(c))
2008 break;
2009 left--;
2010 }
2011 }
2012 if (write && first)
2013 return -EINVAL;
2014 *lenp -= left;
2015 *ppos += *lenp;
2016 return 0;
2017 #undef TMPBUFLEN
2018 }
2019
2020 static int do_proc_doulongvec_minmax(ctl_table *table, int write,
2021 struct file *filp,
2022 void __user *buffer,
2023 size_t *lenp, loff_t *ppos,
2024 unsigned long convmul,
2025 unsigned long convdiv)
2026 {
2027 return __do_proc_doulongvec_minmax(table->data, table, write,
2028 filp, buffer, lenp, ppos, convmul, convdiv);
2029 }
2030
2031 /**
2032 * proc_doulongvec_minmax - read a vector of long integers with min/max values
2033 * @table: the sysctl table
2034 * @write: %TRUE if this is a write to the sysctl file
2035 * @filp: the file structure
2036 * @buffer: the user buffer
2037 * @lenp: the size of the user buffer
2038 * @ppos: file position
2039 *
2040 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2041 * values from/to the user buffer, treated as an ASCII string.
2042 *
2043 * This routine will ensure the values are within the range specified by
2044 * table->extra1 (min) and table->extra2 (max).
2045 *
2046 * Returns 0 on success.
2047 */
2048 int proc_doulongvec_minmax(ctl_table *table, int write, struct file *filp,
2049 void __user *buffer, size_t *lenp, loff_t *ppos)
2050 {
2051 return do_proc_doulongvec_minmax(table, write, filp, buffer, lenp, ppos, 1l, 1l);
2052 }
2053
2054 /**
2055 * proc_doulongvec_ms_jiffies_minmax - read a vector of millisecond values with min/max values
2056 * @table: the sysctl table
2057 * @write: %TRUE if this is a write to the sysctl file
2058 * @filp: the file structure
2059 * @buffer: the user buffer
2060 * @lenp: the size of the user buffer
2061 * @ppos: file position
2062 *
2063 * Reads/writes up to table->maxlen/sizeof(unsigned long) unsigned long
2064 * values from/to the user buffer, treated as an ASCII string. The values
2065 * are treated as milliseconds, and converted to jiffies when they are stored.
2066 *
2067 * This routine will ensure the values are within the range specified by
2068 * table->extra1 (min) and table->extra2 (max).
2069 *
2070 * Returns 0 on success.
2071 */
2072 int proc_doulongvec_ms_jiffies_minmax(ctl_table *table, int write,
2073 struct file *filp,
2074 void __user *buffer,
2075 size_t *lenp, loff_t *ppos)
2076 {
2077 return do_proc_doulongvec_minmax(table, write, filp, buffer,
2078 lenp, ppos, HZ, 1000l);
2079 }
2080
2081
2082 static int do_proc_dointvec_jiffies_conv(int *negp, unsigned long *lvalp,
2083 int *valp,
2084 int write, void *data)
2085 {
2086 if (write) {
2087 if (*lvalp > LONG_MAX / HZ)
2088 return 1;
2089 *valp = *negp ? -(*lvalp*HZ) : (*lvalp*HZ);
2090 } else {
2091 int val = *valp;
2092 unsigned long lval;
2093 if (val < 0) {
2094 *negp = -1;
2095 lval = (unsigned long)-val;
2096 } else {
2097 *negp = 0;
2098 lval = (unsigned long)val;
2099 }
2100 *lvalp = lval / HZ;
2101 }
2102 return 0;
2103 }
2104
2105 static int do_proc_dointvec_userhz_jiffies_conv(int *negp, unsigned long *lvalp,
2106 int *valp,
2107 int write, void *data)
2108 {
2109 if (write) {
2110 if (USER_HZ < HZ && *lvalp > (LONG_MAX / HZ) * USER_HZ)
2111 return 1;
2112 *valp = clock_t_to_jiffies(*negp ? -*lvalp : *lvalp);
2113 } else {
2114 int val = *valp;
2115 unsigned long lval;
2116 if (val < 0) {
2117 *negp = -1;
2118 lval = (unsigned long)-val;
2119 } else {
2120 *negp = 0;
2121 lval = (unsigned long)val;
2122 }
2123 *lvalp = jiffies_to_clock_t(lval);
2124 }
2125 return 0;
2126 }
2127
2128 static int do_proc_dointvec_ms_jiffies_conv(int *negp, unsigned long *lvalp,
2129 int *valp,
2130 int write, void *data)
2131 {
2132 if (write) {
2133 *valp = msecs_to_jiffies(*negp ? -*lvalp : *lvalp);
2134 } else {
2135 int val = *valp;
2136 unsigned long lval;
2137 if (val < 0) {
2138 *negp = -1;
2139 lval = (unsigned long)-val;
2140 } else {
2141 *negp = 0;
2142 lval = (unsigned long)val;
2143 }
2144 *lvalp = jiffies_to_msecs(lval);
2145 }
2146 return 0;
2147 }
2148
2149 /**
2150 * proc_dointvec_jiffies - read a vector of integers as seconds
2151 * @table: the sysctl table
2152 * @write: %TRUE if this is a write to the sysctl file
2153 * @filp: the file structure
2154 * @buffer: the user buffer
2155 * @lenp: the size of the user buffer
2156 * @ppos: file position
2157 *
2158 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2159 * values from/to the user buffer, treated as an ASCII string.
2160 * The values read are assumed to be in seconds, and are converted into
2161 * jiffies.
2162 *
2163 * Returns 0 on success.
2164 */
2165 int proc_dointvec_jiffies(ctl_table *table, int write, struct file *filp,
2166 void __user *buffer, size_t *lenp, loff_t *ppos)
2167 {
2168 return do_proc_dointvec(table,write,filp,buffer,lenp,ppos,
2169 do_proc_dointvec_jiffies_conv,NULL);
2170 }
2171
2172 /**
2173 * proc_dointvec_userhz_jiffies - read a vector of integers as 1/USER_HZ seconds
2174 * @table: the sysctl table
2175 * @write: %TRUE if this is a write to the sysctl file
2176 * @filp: the file structure
2177 * @buffer: the user buffer
2178 * @lenp: the size of the user buffer
2179 * @ppos: pointer to the file position
2180 *
2181 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2182 * values from/to the user buffer, treated as an ASCII string.
2183 * The values read are assumed to be in 1/USER_HZ seconds, and
2184 * are converted into jiffies.
2185 *
2186 * Returns 0 on success.
2187 */
2188 int proc_dointvec_userhz_jiffies(ctl_table *table, int write, struct file *filp,
2189 void __user *buffer, size_t *lenp, loff_t *ppos)
2190 {
2191 return do_proc_dointvec(table,write,filp,buffer,lenp,ppos,
2192 do_proc_dointvec_userhz_jiffies_conv,NULL);
2193 }
2194
2195 /**
2196 * proc_dointvec_ms_jiffies - read a vector of integers as 1 milliseconds
2197 * @table: the sysctl table
2198 * @write: %TRUE if this is a write to the sysctl file
2199 * @filp: the file structure
2200 * @buffer: the user buffer
2201 * @lenp: the size of the user buffer
2202 * @ppos: file position
2203 * @ppos: the current position in the file
2204 *
2205 * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
2206 * values from/to the user buffer, treated as an ASCII string.
2207 * The values read are assumed to be in 1/1000 seconds, and
2208 * are converted into jiffies.
2209 *
2210 * Returns 0 on success.
2211 */
2212 int proc_dointvec_ms_jiffies(ctl_table *table, int write, struct file *filp,
2213 void __user *buffer, size_t *lenp, loff_t *ppos)
2214 {
2215 return do_proc_dointvec(table, write, filp, buffer, lenp, ppos,
2216 do_proc_dointvec_ms_jiffies_conv, NULL);
2217 }
2218
2219 static int proc_do_cad_pid(ctl_table *table, int write, struct file *filp,
2220 void __user *buffer, size_t *lenp, loff_t *ppos)
2221 {
2222 struct pid *new_pid;
2223 pid_t tmp;
2224 int r;
2225
2226 tmp = pid_nr(cad_pid);
2227
2228 r = __do_proc_dointvec(&tmp, table, write, filp, buffer,
2229 lenp, ppos, NULL, NULL);
2230 if (r || !write)
2231 return r;
2232
2233 new_pid = find_get_pid(tmp);
2234 if (!new_pid)
2235 return -ESRCH;
2236
2237 put_pid(xchg(&cad_pid, new_pid));
2238 return 0;
2239 }
2240
2241 #else /* CONFIG_PROC_FS */
2242
2243 int proc_dostring(ctl_table *table, int write, struct file *filp,
2244 void __user *buffer, size_t *lenp, loff_t *ppos)
2245 {
2246 return -ENOSYS;
2247 }
2248
2249 int proc_dointvec(ctl_table *table, int write, struct file *filp,
2250 void __user *buffer, size_t *lenp, loff_t *ppos)
2251 {
2252 return -ENOSYS;
2253 }
2254
2255 int proc_dointvec_bset(ctl_table *table, int write, struct file *filp,
2256 void __user *buffer, size_t *lenp, loff_t *ppos)
2257 {
2258 return -ENOSYS;
2259 }
2260
2261 int proc_dointvec_minmax(ctl_table *table, int write, struct file *filp,
2262 void __user *buffer, size_t *lenp, loff_t *ppos)
2263 {
2264 return -ENOSYS;
2265 }
2266
2267 int proc_dointvec_jiffies(ctl_table *table, int write, struct file *filp,
2268 void __user *buffer, size_t *lenp, loff_t *ppos)
2269 {
2270 return -ENOSYS;
2271 }
2272
2273 int proc_dointvec_userhz_jiffies(ctl_table *table, int write, struct file *filp,
2274 void __user *buffer, size_t *lenp, loff_t *ppos)
2275 {
2276 return -ENOSYS;
2277 }
2278
2279 int proc_dointvec_ms_jiffies(ctl_table *table, int write, struct file *filp,
2280 void __user *buffer, size_t *lenp, loff_t *ppos)
2281 {
2282 return -ENOSYS;
2283 }
2284
2285 int proc_doulongvec_minmax(ctl_table *table, int write, struct file *filp,
2286 void __user *buffer, size_t *lenp, loff_t *ppos)
2287 {
2288 return -ENOSYS;
2289 }
2290
2291 int proc_doulongvec_ms_jiffies_minmax(ctl_table *table, int write,
2292 struct file *filp,
2293 void __user *buffer,
2294 size_t *lenp, loff_t *ppos)
2295 {
2296 return -ENOSYS;
2297 }
2298
2299
2300 #endif /* CONFIG_PROC_FS */
2301
2302
2303 #ifdef CONFIG_SYSCTL_SYSCALL
2304 /*
2305 * General sysctl support routines
2306 */
2307
2308 /* The generic string strategy routine: */
2309 int sysctl_string(ctl_table *table, int __user *name, int nlen,
2310 void __user *oldval, size_t __user *oldlenp,
2311 void __user *newval, size_t newlen)
2312 {
2313 if (!table->data || !table->maxlen)
2314 return -ENOTDIR;
2315
2316 if (oldval && oldlenp) {
2317 size_t bufsize;
2318 if (get_user(bufsize, oldlenp))
2319 return -EFAULT;
2320 if (bufsize) {
2321 size_t len = strlen(table->data), copied;
2322
2323 /* This shouldn't trigger for a well-formed sysctl */
2324 if (len > table->maxlen)
2325 len = table->maxlen;
2326
2327 /* Copy up to a max of bufsize-1 bytes of the string */
2328 copied = (len >= bufsize) ? bufsize - 1 : len;
2329
2330 if (copy_to_user(oldval, table->data, copied) ||
2331 put_user(0, (char __user *)(oldval + copied)))
2332 return -EFAULT;
2333 if (put_user(len, oldlenp))
2334 return -EFAULT;
2335 }
2336 }
2337 if (newval && newlen) {
2338 size_t len = newlen;
2339 if (len > table->maxlen)
2340 len = table->maxlen;
2341 if(copy_from_user(table->data, newval, len))
2342 return -EFAULT;
2343 if (len == table->maxlen)
2344 len--;
2345 ((char *) table->data)[len] = 0;
2346 }
2347 return 1;
2348 }
2349
2350 /*
2351 * This function makes sure that all of the integers in the vector
2352 * are between the minimum and maximum values given in the arrays
2353 * table->extra1 and table->extra2, respectively.
2354 */
2355 int sysctl_intvec(ctl_table *table, int __user *name, int nlen,
2356 void __user *oldval, size_t __user *oldlenp,
2357 void __user *newval, size_t newlen)
2358 {
2359
2360 if (newval && newlen) {
2361 int __user *vec = (int __user *) newval;
2362 int *min = (int *) table->extra1;
2363 int *max = (int *) table->extra2;
2364 size_t length;
2365 int i;
2366
2367 if (newlen % sizeof(int) != 0)
2368 return -EINVAL;
2369
2370 if (!table->extra1 && !table->extra2)
2371 return 0;
2372
2373 if (newlen > table->maxlen)
2374 newlen = table->maxlen;
2375 length = newlen / sizeof(int);
2376
2377 for (i = 0; i < length; i++) {
2378 int value;
2379 if (get_user(value, vec + i))
2380 return -EFAULT;
2381 if (min && value < min[i])
2382 return -EINVAL;
2383 if (max && value > max[i])
2384 return -EINVAL;
2385 }
2386 }
2387 return 0;
2388 }
2389
2390 /* Strategy function to convert jiffies to seconds */
2391 int sysctl_jiffies(ctl_table *table, int __user *name, int nlen,
2392 void __user *oldval, size_t __user *oldlenp,
2393 void __user *newval, size_t newlen)
2394 {
2395 if (oldval && oldlenp) {
2396 size_t olen;
2397
2398 if (get_user(olen, oldlenp))
2399 return -EFAULT;
2400 if (olen) {
2401 int val;
2402
2403 if (olen < sizeof(int))
2404 return -EINVAL;
2405
2406 val = *(int *)(table->data) / HZ;
2407 if (put_user(val, (int __user *)oldval))
2408 return -EFAULT;
2409 if (put_user(sizeof(int), oldlenp))
2410 return -EFAULT;
2411 }
2412 }
2413 if (newval && newlen) {
2414 int new;
2415 if (newlen != sizeof(int))
2416 return -EINVAL;
2417 if (get_user(new, (int __user *)newval))
2418 return -EFAULT;
2419 *(int *)(table->data) = new*HZ;
2420 }
2421 return 1;
2422 }
2423
2424 /* Strategy function to convert jiffies to seconds */
2425 int sysctl_ms_jiffies(ctl_table *table, int __user *name, int nlen,
2426 void __user *oldval, size_t __user *oldlenp,
2427 void __user *newval, size_t newlen)
2428 {
2429 if (oldval && oldlenp) {
2430 size_t olen;
2431
2432 if (get_user(olen, oldlenp))
2433 return -EFAULT;
2434 if (olen) {
2435 int val;
2436
2437 if (olen < sizeof(int))
2438 return -EINVAL;
2439
2440 val = jiffies_to_msecs(*(int *)(table->data));
2441 if (put_user(val, (int __user *)oldval))
2442 return -EFAULT;
2443 if (put_user(sizeof(int), oldlenp))
2444 return -EFAULT;
2445 }
2446 }
2447 if (newval && newlen) {
2448 int new;
2449 if (newlen != sizeof(int))
2450 return -EINVAL;
2451 if (get_user(new, (int __user *)newval))
2452 return -EFAULT;
2453 *(int *)(table->data) = msecs_to_jiffies(new);
2454 }
2455 return 1;
2456 }
2457
2458
2459
2460 #else /* CONFIG_SYSCTL_SYSCALL */
2461
2462
2463 asmlinkage long sys_sysctl(struct __sysctl_args __user *args)
2464 {
2465 static int msg_count;
2466 struct __sysctl_args tmp;
2467 int name[CTL_MAXNAME];
2468 int i;
2469
2470 /* Read in the sysctl name for better debug message logging */
2471 if (copy_from_user(&tmp, args, sizeof(tmp)))
2472 return -EFAULT;
2473 if (tmp.nlen <= 0 || tmp.nlen >= CTL_MAXNAME)
2474 return -ENOTDIR;
2475 for (i = 0; i < tmp.nlen; i++)
2476 if (get_user(name[i], tmp.name + i))
2477 return -EFAULT;
2478
2479 /* Ignore accesses to kernel.version */
2480 if ((tmp.nlen == 2) && (name[0] == CTL_KERN) && (name[1] == KERN_VERSION))
2481 goto out;
2482
2483 if (msg_count < 5) {
2484 msg_count++;
2485 printk(KERN_INFO
2486 "warning: process `%s' used the removed sysctl "
2487 "system call with ", current->comm);
2488 for (i = 0; i < tmp.nlen; i++)
2489 printk("%d.", name[i]);
2490 printk("\n");
2491 }
2492 out:
2493 return -ENOSYS;
2494 }
2495
2496 int sysctl_string(ctl_table *table, int __user *name, int nlen,
2497 void __user *oldval, size_t __user *oldlenp,
2498 void __user *newval, size_t newlen)
2499 {
2500 return -ENOSYS;
2501 }
2502
2503 int sysctl_intvec(ctl_table *table, int __user *name, int nlen,
2504 void __user *oldval, size_t __user *oldlenp,
2505 void __user *newval, size_t newlen)
2506 {
2507 return -ENOSYS;
2508 }
2509
2510 int sysctl_jiffies(ctl_table *table, int __user *name, int nlen,
2511 void __user *oldval, size_t __user *oldlenp,
2512 void __user *newval, size_t newlen)
2513 {
2514 return -ENOSYS;
2515 }
2516
2517 int sysctl_ms_jiffies(ctl_table *table, int __user *name, int nlen,
2518 void __user *oldval, size_t __user *oldlenp,
2519 void __user *newval, size_t newlen)
2520 {
2521 return -ENOSYS;
2522 }
2523
2524 #endif /* CONFIG_SYSCTL_SYSCALL */
2525
2526 /*
2527 * No sense putting this after each symbol definition, twice,
2528 * exception granted :-)
2529 */
2530 EXPORT_SYMBOL(proc_dointvec);
2531 EXPORT_SYMBOL(proc_dointvec_jiffies);
2532 EXPORT_SYMBOL(proc_dointvec_minmax);
2533 EXPORT_SYMBOL(proc_dointvec_userhz_jiffies);
2534 EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
2535 EXPORT_SYMBOL(proc_dostring);
2536 EXPORT_SYMBOL(proc_doulongvec_minmax);
2537 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
2538 EXPORT_SYMBOL(register_sysctl_table);
2539 EXPORT_SYMBOL(sysctl_intvec);
2540 EXPORT_SYMBOL(sysctl_jiffies);
2541 EXPORT_SYMBOL(sysctl_ms_jiffies);
2542 EXPORT_SYMBOL(sysctl_string);
2543 EXPORT_SYMBOL(unregister_sysctl_table);
kernel source for telekom puls (alcatel/mediatek)