Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 3 Sep 2013 17:02:43 -0600
From: Vincent Danen <vdanen@...hat.com>
To: oss-security@...ts.openwall.com, kseifried@...hat.com
Cc: Jonas Meurer <jonas@...esources.org>, contribute@...ios.org
Subject: Re: CVE request: unauthorized host/service views
 displayed in servicegroup view

* [2013-09-03 13:41:03 -0600] Kurt Seifried wrote:

>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>On 08/30/2013 09:47 AM, Jonas Meurer wrote:
>> Any news on that?
>>
>> I still believe that there's a misunderstanding. I still consider
>> the bug I found as security relevant. Hostnames are leaked to
>> unauthorized nagios-cgi users.
>>
>> In case that you don't agree with what I've written below, please
>> explain.
>>
>> Honestly, I cannot believe that it was meant as a _feature_ by
>> nagios devs that _all_ hostnames are displayed for _all_ users,
>> regardless whether they're listed in contacts/contactgroups.
>>
>> I don't consider this issue too important, but still a CVE would
>> be appropriative in my opinion. It should be fixed in future
>> uploads of nagios3 to the major distributions. Most nagios admins
>> might be unaware of this issue.
>>
>> Kind regards, jonas
>>
>>
>> Am 2013-08-04 02:40, schrieb Jonas Meurer:
>>> Hello,
>>>
>>> sorry, I'm on holidays and cannot work on this issue for the next
>>> two weeks. But I think that there is a missunderstanding. See my
>>> short comment below.
>>>
>>> Am 02.08.2013 19:27, schrieb Vincent Danen:
>>>> * [2013-07-10 17:17:08 +0200] Jonas Meurer wrote:
>>>>
>>>>> Hello,
>>>>>
>>>>> Am 2013-07-08 20:16, schrieb Kurt Seifried:
>>>>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
>>>>>>
>>>>>> On 06/26/2013 01:42 PM, Kurt Seifried wrote:
>>>>>>> On 06/26/2013 12:36 PM, Vincent Danen wrote:
>>>>>>>> I don't believe a CVE has been assigned to this issue
>>>>>>>> yet.
>>>>>>>
>>>>>>>> It was reported that Nagios 3.4.4 at least, and
>>>>>>>> possibly earlier versions, would allow users with
>>>>>>>> access to Nagios to obtain full access to the
>>>>>>>> servicegroup overview, even if they are not authorized
>>>>>>>> to view all of the systems (not configured for this
>>>>>>>> ability in the authorized_for_* configuration option).
>>>>>>>> This includes the servicegroup overview, summary, and
>>>>>>>> grid.
>>>>>>>
>>>>>>>> Provided the user has access to view some services,
>>>>>>>> they will be able to see all services (including those
>>>>>>>> they should not see). Note that the user in question
>>>>>>>> must have access to some services and must have access
>>>>>>>> to Nagios to begin with.
>>>>>>>
>>>>>>>> This has not yet been corrected upstream.
>>>>>>>
>>>>>>>> References:
>>>>>>>
>>>>>>>> http://www.mail-archive.com/nagios-users@lists.sourceforge.net/msg39749.html
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>>
>>>>>>>>
>http://tracker.nagios.org/view.php?id=456
>>>>>>>> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=714171
>>>>>>>>
>>>>>>>>
>https://bugzilla.redhat.com/show_bug.cgi?id=978531
>>>>>>>
>>>>>>>
>>>>>>>> Thanks.
>>>>>>>
>>>>>>> Please use CVE-2013-2214 for this issue.
>>>>>>
>>>>>> It appears there are may be some problems with this issue,
>>>>>> potentially this may have been a bad configuration and not
>>>>>> a source code based problem, however we haven't been able
>>>>>> to confirm it yet. I've also not been able to contact
>>>>>> upstream about this easily (no security@ address, if anyone
>>>>>> know whom to forward this to, please let me know, thanks.
>>>>>
>>>>> I'm wondering why you fail to reproduce this issue. I posted
>>>>> some details regarding my setup at the Nagios Tracker:
>>>>> http://tracker.nagios.org/view.php?id=456
>>>>>
>>>>> Unfortunately Nagios upstream sometimes rather unresponsive.
>>>>> At least that's what I observed.
>>>>>
>>>>> Please let me know if you need any further details regarding
>>>>> the bug or advice on how to reproduce it.
>>>>
>>>> To close the loop on this, the CVE should probably be
>>>> rejected. According to upstream, this is done by design.  One
>>>> of our users noted it in our bugzilla:
>>>>
>>>> https://bugzilla.redhat.com/show_bug.cgi?id=978531#c11
>>>>
>>>> He has a thorough explanation, but the bottom line is this
>>>> seems to be by design, as noted in the changelog:
>>>>
>>>> http://www.nagios.org/projects/nagioscore/history/core-3x
>>>>
>>>> * Users can now see hostgroups and servicegroups that contain
>>>> at least one host or service they are authorized for, instead
>>>> of having to be authorized for them all (Ethan Galstad)
>>>
>>> As I understand this changelog entry, it means the following:
>>>
>>> Hostgroups and servicegroups are listed with all the
>>> _authorized_ members if the user is authorized to see at least
>>> one member.
>>>
>>> To me it doesn't mean the following (which was the case without
>>> my patch):
>>>
>>> Servicegroups are listed with all members (regardless wether
>>> authorized or unauthorized) if the user is authorized to see at
>>> least one member.
>>>
>>> Another argument for my point of view is that the nagios
>>> maintainers (silently) accepted my patch (at least if I remember
>>> correctly, it has been incorporated into the upstream development
>>> repository). Unfortunately there's still not one single statement
>>> from upstream about the issue, that I'm aware of.
>>>
>>>> I suspect this CVE should be rejected as this is done by
>>>> design.
>>>
>>> Like explained above, I disagree with this suggestion :)
>>>
>>> Kind regards, jonas
>
>https://bugzilla.redhat.com/show_bug.cgi?id=978531#c11
>
>So as I understand it, NOTABUG, upstream documented it, etc.

Until and unless upstream acknowledges that this isn't by design (which
the changelog entry most definitely indicates it _is_ by design) then I
don't think there's anything we can do.

I mean, if someone wants to shoot themselves in the foot and document it
as a feature, who are we to say otherwise?  We may not agree with it,
but it's a documented feature (deliberately changed), so we can't just
very well call it a security flaw because we don't like the new
behaviour.

Having said that, it would be fantastic if the folks at Nagios would
speak up and put this to rest.  It could be a design decision that went
too lax (thus being a flaw, exposing more than intended), or this might
actually be what was meant.  Until upstream says one way or another, I
would consider this not a security flaw.

-- 
Vincent Danen / Red Hat Security Response Team 

Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.