[Snort-sigs] LDAPv3 with simple authentication

FOULDE Damien damien.foulde at ...4205...
Thu Dec 29 10:36:37 EST 2016


Hello,

 

The previous message is currently waiting for a moderator approval.

As it’s not released for the moment, I registered to the maillinglist and
here it is.

 

Regards,

 

Damien

 

De : FOULDE Damien 
Envoyé : mercredi 21 décembre 2016 10:51
À : snort-sigs at lists.sourceforge.net
Objet : RE: LDAPv3 with simple authentication

 

Hello,

 

Any ideas / suggestions / advices will be greatly appreciated regarding this
question.

 

In the meantime, here’s a working signature without fully decoding the BER
data :

alert tcp any any -> any 389 (sid:1000000; gid:1;
flow:established,to_server; content:"|30|"; depth:1; content:"|02|";
distance:1; within:127; content:"|60|"; distance:1; within:5; content:"|02
01 03 04|"; fast_pattern; distance:1; within:127; content:"|80|";
distance:1; within:127; content:!"|02 01 03 04 00 a3|"; offset:7; depth:257;
msg:"LDAPv3 simple Authentication"; classtype:policy-violation; rev:1; )

 

It would be great if it could be reviewed by Talos.

I can provide a packet capture if needed.

 

Regards,

 

Damien

 

De : FOULDE Damien [mailto:damien.foulde at ...4205...] 
Envoyé : lundi 19 décembre 2016 12:39
À : snort-sigs at lists.sourceforge.net
Objet : [Snort-sigs] LDAPv3 with simple authentication

 

Hello,

 

We need to write a signature to match on LDAPv3 with simple authentication.

LDAPv3 is described in the RFC 2251 through Abstract Syntax Notation 1
(ASN.1) and encoded through a subset of Basic Encoding Rules (BER) in the
packets.

You may have a look to this great website
http://www.selfadsi.org/ldap.htm#Frame to have a quick look over the
encoding.

https://en.wikipedia.org/wiki/X.690#BER_encoding is also a good source of
information.

As you should have seen the length can be encoded in a short or long form.

When the short form is used the MSB is set to 0 and the 7 remaining bits are
used to encode the length directly from 0 to 127.

Using the byte_jump function we should be able to jump to the next encoded
data.

When the long form is used the MSB is set to 1 and the 7 remaining bits are
used to encode the number of bytes that follow from 1 to 126 which will
contains the actual length.

Using byte_extract and byte_jump functions with bitmask we should be able to
jump to the next encoded data.

Before reaching the point where the LDAPv3 authentication is set to simple
(encoded to 0) or sasl (encoded to 3) there’re 5 short or long length bytes.

Is there a way through the subset of snort packet dissection functions to
match on this without writing 32 (2^5) different signatures to match all
short / long possibilities ?

The BER encoding is also used to encode SNMP, the same kind of issue may
have been seen there also.

 

Thank you for your help,

 

Damien

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.snort.org/pipermail/snort-sigs/attachments/20161229/35b7a8fb/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5355 bytes
Desc: not available
URL: <https://lists.snort.org/pipermail/snort-sigs/attachments/20161229/35b7a8fb/attachment.bin>


More information about the Snort-sigs mailing list