-a task,always audits everything

-a task,always audits everything

[Matthew Booth]

[-- Attachment #1.1: Type: text/plain, Size: 547 bytes --]

I added the following to audit.rules in RHEL4 U4 x86_64:

-a task,always

I expected this to have a similar effect to auditing the fork, vfork and
clone system calls. However, it seems to either audit everything itself,
or cause all the filtering to be removed from my other audit rules
(which audit pretty much everything).

Is this expected behaviour, or a bug?

Thanks,

Matt
-- 
Red Hat, Global Professional Services

M:       +44 (0)7977 267231
GPG ID:  D33C3490
GPG FPR: 3733 612D 2D05 5458 8A8A 1600 3441 EA19 D33C 3490

[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: -a task,always audits everything

[Steve Grubb]

On Wednesday 21 February 2007 06:14:26 Matthew Booth wrote:
> I added the following to audit.rules in RHEL4 U4 x86_64:
>
> -a task,always
>
> I expected this to have a similar effect to auditing the fork, vfork and
> clone system calls.

Task auditing and syscall auditing are a little different. Task filters would 
most likely be used a field (-F) option. What the above is doing is marking 
each task auditable as its created.

> However, it seems to either audit everything itself, 

As given, I think so. What happens is that with no fields, the 
audit_filter_rules() function in the kernel avoids the switch case and winds 
up in another branch which is for situations like:

auditctl -a always,exit -s open

In that case, you would want to audit any call to open. So, what happens is 
each task is marked auditable as its created since there are no field 
qualifiers to discriminate with.

> or cause all the filtering to be removed from my other audit rules (which
> audit pretty much everything). 

Audit is a first match wins system.

> Is this expected behaviour, or a bug?

I think this is expected.

-Steve