On 9/16/10 4:24 PM, Mr Dash Four wrote:
> 
>> As I announced yesterday on the Development List, I've concluded that
>> the direction in which blacklisting has been going since 4.4.12 is not
>> appropriate. The current plan for 4.4.13 is to return blacklisting to
>> the way it was in 4.4.11; basically, support for the OPTIONS column will
>> be removed.
>>
>> As I've thought more about this issue, I have realized that the
>> fundamental problem with extending the 4.4.11 support is that while
>> blacklisting is a security-related concept, in 4.4.11 and earlier
>> blacklisting support has been associated with interfaces or host groups.
>> A much more natural approach is to associate blacklisting with zones.
>>   
> I concur! Also, you could always add support (if it doesn't already 
> exist - I am no Shorewall expert) to include individual interfaces from 
> each zone, like "net:eth1" for example.
> 
>> I have prototyped the following proposal and it seems to work correctly.
>>
>> 1) The OPTIONS column of the blacklist file will be restored to the way
>>    that it has been in the recent 4.4.13 Betas. The preferred keywords
>>    are 'src' and 'dst' rather than 'from' and 'to' although the latter
>>    would also be supported. Duplicates ignored with a warning.
>>   
> Makes sense and it was the way I preferred - it makes it consistent with 
> the rest (rules etc).
> 
>> 1) Allow 'blacklist' in the OPTIONS, IN_OPTIONS and OUT_OPTIONS columns
>>    of the zones file.
>>
>>      Specifying 'blacklist' in OPTIONS, is equivalent to specifying
>>      it in both IN_OPTIONS and OUT_OPTIONS.
>>   
> So, if I specify 'blacklist' in my net zone then it would automatically 
> be bidirectional provided I have "+backlist-nets src,dst" in my 
> blacklist file, right? If so, that would be perfect.

That's the plan.

> 
>> 4)  In the interfaces file:
>>
>>      If 'blacklist' is specified with no ZONE, a warning is given
>>      and the option is ignored.
>>
>>      If 'blacklist' is specified with a ZONE, it is equivalent to
>>      specifying 'blacklist' in the IN_OPTIONS column of the
>>      corresponding entry in the zones file.
>>
>>      Currently, incoming blacklisting is performed as a first step
>>      on incoming packets on those interfaces having the 'blacklist'
>>      option. With the new approach, blacklisting would occur after
>>      the packet has been processed against the interface-related
>>      options such as 'nosmurfs', 'tcpflags', etc.
>>   
> Is there any reason for this? For me it would not make sense to do ANY 
> processing (ddos etc) if this packet is/has failed the blacklist test.

It's a consequence of the order in which the ruleset is constructed.
Interface-oriented filtering is generated well before zone-oriented
filtering. During optimization, interface-oriented filtering can be
moved to the front of a zone-oriented chain. It's something that could
be overcome but it's not trivial.

-Tom
-- 
Tom Eastep        \ When I die, I want to go like my Grandfather who
Shoreline,         \ died peacefully in his sleep. Not screaming like
Washington, USA     \ all of the passengers in his car
http://shorewall.net \________________________________________________

Attachment: signature.asc
Description: OpenPGP digital signature

------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
Shorewall-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/shorewall-devel

Reply via email to