[Snort-sigs] A question on ethernet padding

James Lay jlay at ...3266...
Thu Jan 23 15:26:14 EST 2014

On 2014-01-23 13:20, Jeremy Hoel wrote:
> So there you go.. I was trying various offsets and depths and didn't
> seem to get it.  But I'll try that.  Thanks!
> BTW - Should that be part of the rule?  Since you wouldn't want those
> to fire if they had 0 data?
> On Thu, Jan 23, 2014 at 8:17 PM, James Lay <jlay at ...3266...> 
> wrote:
>> On 2014-01-23 12:54, Jeremy Hoel wrote:
>>> I was wondering kind of the same question.. in regards to those new
>>> ICMP rules.  NetApps doing have any ICMP data, just the main 
>>> requests,
>>> but there seems to always be 10 bytes |00| in what wireshark calls
>>> padding, and I'm curious if I can write the rule around that.
>>> On Thu, Jan 23, 2014 at 4:07 PM, James Lay 
>>> <jlay at ...3266...>
>>> wrote:
>>>> Does snort treat ethernet padding as data?  Wireshark shows that I 
>>>> have
>>>> 1 byte of data in a packet after my ethernet and ip headers.  My
>>>> ethernet header, normally 14 bytes, includes 17 bytes of Padding.  
>>>> Does
>>>> snort consider the padding as data?  Trying to figure out what 
>>>> offset
>>>> and depth to use on this rule.  Hope I'm explaining this 
>>>> well..thanks
>>>> all.
>>>> James
>> An end around around to NOT see these can be to add dsize:>1; to 
>> your
>> rule...should nuke out these zero data pings.
>> James

Yea..I've been making specific rules to match standard ping types that 
are anomalous, and then a catch all rule.  Here's my catch all so far:

alert icmp any any -> any any (msg:"Unusual PING detected"; icode:0; 
itype:8; fragbits:!M; ttl:>10; dsize:>5; 
content:!"0123456789abcdefghijklmnopqrstuv"; depth:32; 
content:!"abcdefghijklmnopqrstuvwabcdefghi"; depth:32; 
classtype:bad-unknown; sid:10000119; rev:4;)

It's been a neat exercise seeing how I can hone it to fire on just 
what's weird, not what's usual.


More information about the Snort-sigs mailing list