1AUDITCTL:(8) System Administration Utilities AUDITCTL:(8)
2
6 auditctl - a utility to assist controlling the kernel's audit system
7
9 auditctl [options]
10
12 The auditctl program is used to configure kernel options related to
13 auditing, to see status of the configuration, and to load discretionary
14 audit rules.
15
17 -b backlog
18 Set max number (limit) of outstanding audit buffers allowed
19 (Kernel Default=64) If all buffers are full, the failure flag is
20 consulted by the kernel for action.
21 --backlog_wait_time wait_time
23 Set the time for the kernel to wait (Kernel Default 60*HZ) when
24 the backlog limit is reached before queuing more audit events to
25 be transferred to auditd. The number must be greater than or
26 equal to zero and less that 10 times the default value.
27 -c Continue loading rules in spite of an error. This summarizes the
29 results of loading the rules. The exit code will not be success
30 if any rule fails to load.
31 -D Delete all rules and watches. This can take a key option (-k),
33 too.
34 -e [0..2]
36 Set enabled flag. When 0 is passed, this can be used to tempo‐
37 rarily disable auditing. When 1 is passed as an argument, it
38 will enable auditing. To lock the audit configuration so that it
39 can't be changed, pass a 2 as the argument. Locking the configu‐
40 ration is intended to be the last command in audit.rules for
41 anyone wishing this feature to be active. Any attempt to change
42 the configuration in this mode will be audited and denied. The
43 configuration can only be changed by rebooting the machine.
44 -f [0..2]
46 Set failure mode 0=silent 1=printk 2=panic. This option lets you
47 determine how you want the kernel to handle critical errors.
48 Example conditions where this mode may have an effect includes:
49 transmission errors to userspace audit daemon, backlog limit
50 exceeded, out of kernel memory, and rate limit exceeded. The
51 default value is 1. Secure environments will probably want to
52 set this to 2.
53 -h Help
55 -i When given by itself, ignore errors when reading rules from a
57 file. This causes auditctl to always return a success exit code.
58 If passed as an argument to -s then it gives an interpretation
59 of the numbers to human readable words if possible.
60 --loginuid-immutable
62 This option tells the kernel to make loginuids unchangeable once
63 they are set. Changing loginuids requires CAP_AUDIT_CONTROL. So,
64 its not something that can be done by unprivileged users. Set‐
65 ting this makes loginuid tamper-proof, but can cause some prob‐
66 lems in certain kinds of containers.
67 -q mount-point,subtree
69 If you have an existing directory watch and bind or move mount
70 another subtree in the watched subtree, you need to tell the
71 kernel to make the subtree being mounted equivalent to the
72 directory being watched. If the subtree is already mounted at
73 the time the directory watch is issued, the subtree is automati‐
74 cally tagged for watching. Please note the comma separating the
75 two values. Omitting it will cause errors.
76 -r rate
78 Set limit in messages/sec (0=none). If this rate is non-zero and
79 is exceeded, the failure flag is consulted by the kernel for
80 action. The default value is 0.
81 --reset-lost
83 Reset the lost record counter shown by the status command.
84 -R file
86 Read rules from a file. The rules must be 1 per line and in the
87 order that they are to be executed in. The rule file must be
88 owned by root and not readable by other users or it will be
89 rejected. The rule file may have comments embedded by starting
90 the line with a '#' character. Rules that are read from a file
91 are identical to what you would type on a command line except
92 they are not preceded by auditctl (since auditctl is the one
93 executing the file) and you would not use shell escaping since
94 auditctl is reading the file instead of bash.
95 -t Trim the subtrees after a mount command.
97
99 -l List all rules 1 per line. Two more options may be given to this
100 command. You can give either a key option (-k) to list rules
101 that match a key or a (-i) to have a0 through a3 interpreted to
102 help determine the syscall argument values are correct .
103 -m text
105 Send a user space message into the audit system. This can only
106 be done if you have CAP_AUDIT_WRITE capability (normally the
107 root user has this). The resulting event will be the USER type.
108 -s Report the kernel's audit subsystem status. It will tell you the
110 in-kernel values that can be set by -e, -f, -r, and -b options.
111 The pid value is the process number of the audit daemon. Note
112 that a pid of 0 indicates that the audit daemon is not running.
113 The lost entry will tell you how many event records that have
114 been discarded due to the kernel audit queue overflowing. The
115 backlog field tells how many event records are currently queued
116 waiting for auditd to read them. This option can be followed by
117 the -i to get a couple fields interpreted.
118 -v Print the version of auditctl.
120
123 -a [list,action|action,list]
124 Append rule to the end of list with action. Please note the
125 comma separating the two values. Omitting it will cause errors.
126 The fields may be in either order. It could be list,action or
127 action,list. The following describes the valid list names:
128 task Add a rule to the per task list. This rule list is
130 used only at the time a task is created -- when
131 fork() or clone() are called by the parent task.
132 When using this list, you should only use fields
133 that are known at task creation time, such as the
134 uid, gid, etc.
135 exit Add a rule to the syscall exit list. This list is
137 used upon exit from a system call to determine if an
138 audit event should be created.
139 user Add a rule to the user message filter list. This
141 list is used by the kernel to filter events origi‐
142 nating in user space before relaying them to the
143 audit daemon. It should be noted that the only
144 fields that are valid are: uid, auid, gid, pid,
145 subj_user, subj_role, subj_type, subj_sen, subj_clr,
146 msgtype, and executable name. All other fields will
147 be treated as non-matching. It should be understood
148 that any event originating from user space from a
149 process that has CAP_AUDIT_WRITE will be recorded
150 into the audit trail. This means that the most
151 likely use for this filter is with rules that have
152 an action of never since nothing has to be done to
153 allow events to be recorded.
154 exclude Add a rule to the event type exclusion filter list.
156 This list is used to filter events that you do not
157 want to see. For example, if you do not want to see
158 any avc messages, you would using this list to
159 record that. Events can be excluded by process ID,
160 user ID, group ID, login user ID, message type, sub‐
161 ject context, or executable name. The action is
162 ignored and uses its default of "never".
163 filesystem Add a rule that will be applied to a whole filesys‐
165 tem. The filesystem must be identified with a fstype
166 field. Normally this filter is used to exclude any
167 events for a whole filesystem such as tracefs or
168 debugfs.
169 The following describes the valid actions for the rule:
171 never No audit records will be generated. This can be used
173 to suppress event generation. In general, you want
174 suppressions at the top of the list instead of the
175 bottom. This is because the event triggers on the
176 first matching rule.
177 always Allocate an audit context, always fill it in at
179 syscall entry time, and always write out a record at
180 syscall exit time.
181 -A list,action
183 Add rule to the beginning list with action.
184 -C [f=f | f!=f]
186 Build an inter-field comparison rule: field, operation, field.
187 You may pass multiple comparisons on a single command line. Each
188 one must start with -C. Each inter-field equation is anded with
189 each other as well as equations starting with -F to trigger an
190 audit record. There are 2 operators supported - equal, and not
191 equal. Valid fields are:
192 auid, uid, euid, suid, fsuid, obj_uid; and gid, egid, sgid,
194 fsgid, obj_gid
195 The two groups of uid and gid cannot be mixed. But any compari‐
197 son within the group can be made. The obj_uid/gid fields are
198 collected from the object of the event such as a file or direc‐
199 tory.
200 -d list,action
203 Delete rule from list with action. The rule is deleted only if
204 it exactly matches syscall name(s) and every field name and
205 value.
206 -F [n=v | n!=v | n<v | n>v | n<=v | n>=v | n&v | n&=v]
208 Build a rule field: name, operation, value. You may have up to
209 64 fields passed on a single command line. Each one must start
210 with -F. Each field equation is anded with each other (as well
211 as equations starting with -C) to trigger an audit record. There
212 are 8 operators supported - equal, not equal, less than, greater
213 than, less than or equal, and greater than or equal, bit mask,
214 and bit test respectively. Bit test will "and" the values and
215 check that they are equal, bit mask just "ands" the values.
216 Fields that take a user ID may instead have the user's name; the
217 program will convert the name to user ID. The same is true of
218 group names. Valid fields are:
219 a0, a1, a2, a3
221 Respectively, the first 4 arguments to a syscall.
222 Note that string arguments are not supported. This
223 is because the kernel is passed a pointer to the
224 string. Triggering on a pointer address value is not
225 likely to work. So, when using this, you should only
226 use on numeric values. This is most likely to be
227 used on platforms that multiplex socket or IPC oper‐
228 ations.
229 arch The CPU architecture of the syscall. The arch can be
231 found doing 'uname -m'. If you do not know the arch
232 of your machine but you want to use the 32 bit
233 syscall table and your machine supports 32 bit, you
234 can also use b32 for the arch. The same applies to
235 the 64 bit syscall table, you can use b64. In this
236 way, you can write rules that are somewhat arch
237 independent because the family type will be auto
238 detected. However, syscalls can be arch specific and
239 what is available on x86_64, may not be available on
240 ppc. The arch directive should precede the -S option
241 so that auditctl knows which internal table to use
242 to look up the syscall numbers.
243 auid The original ID the user logged in with. Its an
245 abbreviation of audit uid. Sometimes its referred to
246 as loginuid. Either the user account text or number
247 may be used.
248 devmajor Device Major Number
250 devminor Device Minor Number
252 dir Full Path of Directory to watch. This will place a
254 recursive watch on the directory and its whole sub‐
255 tree. It can only be used on exit list. See "-w".
256 egid Effective Group ID. May be numeric or the groups
258 name.
259 euid Effective User ID. May be numeric or the user
261 account name.
262 exe Absolute path to application that while executing
264 this rule will apply to. It supports = and != opera‐
265 tors. Note that you can only use this once for each
266 rule.
267 exit Exit value from a syscall. If the exit code is an
269 errno, you may use the text representation, too.
270 fsgid Filesystem Group ID. May be numeric or the groups
272 name.
273 fsuid Filesystem User ID. May be numeric or the user
275 account name.
276 filetype The target file's type. Can be either file, dir,
278 socket, link, character, block, or fifo.
279 gid Group ID. May be numeric or the groups name.
281 inode Inode Number
283 key This is another way of setting a filter key. See
285 discussion above for -k option.
286 msgtype This is used to match the event's record type. It
288 should only be used on the exclude or user filter
289 lists.
290 obj_uid Object's UID
292 obj_gid Object's GID
294 obj_user Resource's SE Linux User
296 obj_role Resource's SE Linux Role
298 obj_type Resource's SE Linux Type
300 obj_lev_low Resource's SE Linux Low Level
302 obj_lev_high
304 Resource's SE Linux High Level
305 path Full Path of File to watch. It can only be used on
307 exit list.
308 perm Permission filter for file operations. See "-p". It
310 can only be used on exit list. You can use this
311 without specifying a syscall and the kernel will
312 select the syscalls that satisfy the permissions
313 being requested.
314 pers OS Personality Number
316 pid Process ID
318 ppid Parent's Process ID
320 saddr_fam Address family number as found in
322 /usr/include/bits/socket.h. For example, IPv4 would
323 be 2 and IPv6 would be 10.
324 sessionid User's login session ID
326 subj_user Program's SE Linux User
328 subj_role Program's SE Linux Role
330 subj_type Program's SE Linux Type
332 subj_sen Program's SE Linux Sensitivity
334 subj_clr Program's SE Linux Clearance
336 sgid Saved Group ID. See getresgid(2) man page.
338 success If the exit value is >= 0 this is true/yes otherwise
340 its false/no. When writing a rule, use a 1 for
341 true/yes and a 0 for false/no
342 suid Saved User ID. See getresuid(2) man page.
344 uid User ID. May be numeric or the user account name.
346 -k key Set a filter key on an audit rule. The filter key is an arbi‐
348 trary string of text that can be up to 31 bytes long. It can
349 uniquely identify the audit records produced by a rule. Typical
350 use is for when you have several rules that together satisfy a
351 security requirement. The key value can be searched on with
352 ausearch so that no matter which rule triggered the event, you
353 can find its results. The key can also be used on delete all
354 (-D) and list rules (-l) to select rules with a specific key.
355 You may have more than one key on a rule if you want to be able
356 to search logged events in multiple ways or if you have an
357 auditd plugin that uses a key to aid its analysis.
358 -p [r|w|x|a]
360 Describe the permission access type that a file system watch
361 will trigger on. r=read, w=write, x=execute, a=attribute change.
362 These permissions are not the standard file permissions, but
363 rather the kind of syscall that would do this kind of thing. The
364 read & write syscalls are omitted from this set since they would
365 overwhelm the logs. But rather for reads or writes, the open
366 flags are looked at to see what permission was requested.
367 -S [Syscall name or number|all]
369 Any syscall name or number may be used. The word 'all' may also
370 be used. If the given syscall is made by a program, then start
371 an audit record. If a field rule is given and no syscall is
372 specified, it will default to all syscalls. You may also specify
373 multiple syscalls in the same rule by using multiple -S options
374 in the same rule. Doing so improves performance since fewer
375 rules need to be evaluated. Alternatively, you may pass a comma
376 separated list of syscall names. If you are on a bi-arch system,
377 like x86_64, you should be aware that auditctl simply takes the
378 text, looks it up for the native arch (in this case b64) and
379 sends that rule to the kernel. If there are no additional arch
380 directives, IT WILL APPLY TO BOTH 32 & 64 BIT SYSCALLS. This can
381 have undesirable effects since there is no guarantee that any
382 syscall has the same number on both 32 and 64 bit interfaces.
383 You will likely want to control this and write 2 rules, one with
384 arch equal to b32 and one with b64 to make sure the kernel finds
385 the events that you intend. See the arch field discussion for
386 more info.
387 -w path
389 Insert a watch for the file system object at path. You cannot
390 insert a watch to the top level directory. This is prohibited by
391 the kernel. Wildcards are not supported either and will generate
392 a warning. The way that watches work is by tracking the inode
393 internally. If you place a watch on a file, its the same as
394 using the -F path option on a syscall rule. If you place a watch
395 on a directory, its the same as using the -F dir option on a
396 syscall rule. The -w form of writing watches is for backwards
397 compatibility and the syscall based form is more expressive.
398 Unlike most syscall auditing rules, watches do not impact per‐
399 formance based on the number of rules sent to the kernel. The
400 only valid options when using a watch are the -p and -k. If you
401 need to anything fancy like audit a specific user accessing a
402 file, then use the syscall auditing form with the path or dir
403 fields. See the EXAMPLES section for an example of converting
404 one form to another.
405 -W path
407 Remove a watch for the file system object at path. The rule must
408 match exactly. See -d discussion for more info.
409
411 Syscall rules get evaluated for each syscall for every program. If you
412 have 10 syscall rules, every program on your system will delay during a
413 syscall while the audit system evaluates each rule. Too many syscall
414 rules will hurt performance. Try to combine as many as you can whenever
415 the filter, action, key, and fields are identical. For example:
416 auditctl -a always,exit -S openat -F success=0
418 auditctl -a always,exit -S truncate -F success=0
419 could be re-written as one rule:
421 auditctl -a always,exit -S openat -S truncate -F success=0
423 Also, try to use file system auditing wherever practical. This improves
425 performance. For example, if you were wanting to capture all failed
426 opens & truncates like above, but were only concerned about files in
427 /etc and didn't care about /usr or /sbin, its possible to use this
428 rule:
429 auditctl -a always,exit -S openat -S truncate -F dir=/etc -F success=0
431 This will be higher performance since the kernel will not evaluate it
433 each and every syscall. It will be handled by the filesystem auditing
434 code and only checked on filesystem related syscalls.
435
437 To see all syscalls made by a specific program:
438 auditctl -a always,exit -S all -F pid=1005
440 To see files opened by a specific user:
442 auditctl -a always,exit -S openat -F auid=510
444 To see unsuccessful openat calls:
446 auditctl -a always,exit -S openat -F success=0
448 To watch a file for changes (2 ways to express):
450 auditctl -w /etc/shadow -p wa
452 auditctl -a always,exit -F path=/etc/shadow -F perm=wa
453 To recursively watch a directory for changes (2 ways to express):
455 auditctl -w /etc/ -p wa
457 auditctl -a always,exit -F dir=/etc/ -F perm=wa
458 To see if an admin is accessing other user's files:
460 auditctl -a always,exit -F dir=/home/ -F uid=0 -C auid!=obj_uid
462
465 /etc/audit/audit.rules /etc/audit/audit-stop.rules
466
469 audit.rules(7), auditd(8).
470
473 Steve Grubb
474Red Hat Aug 2018 AUDITCTL:(8)