On 5/17/06, Marcel Ho= ltmann <marcel@holtmann.org> wrote:

Hi Ronnie,=

> Since the patch is 4 years old it does not apply to ethereal s= vn any
> more since too many internal APIs have changed.
>= I am however massaging and refactoring it for inclusion and have
> i= ncluded some of the protocol stack into ethereal already.

I would li= ke to to see full Bluetooth protocol support in Ethereal and
try to help you whereever possible.

Thanks. I = need as many captures as possible. I also need bugreports and suggestions h= ow to "prettify" the dissection so that the important information= in a packet is easy to find when parsing the trace.

It would also be great with someone adding some pages for the vario= us protocols at wiki.ethereal.co= m with links to protocol standards and example captur= e files.

> Et= hereal reads the files generated by some tool called 'hcidump' (i am
>= ; not a bluetooth guy, im a san/nas guy so im not re= ad up on your
> tools)

The program hcidump supports the native hcidump file= format and also the
BTSnoop file format. I might need to send you a pat= ch for it at some
point.

> I am currently refactoring the L2CA= P layer and would really benefit
> from having some more example captures to work and verify with bef= ore
> i check this layer in.

Post a pre-patch to this mailing = list and I am going to have a look at
it.

I have ch= ecked into current svn of ethereal dissectors up to and including the l2cap= layer.
No protocols above the l2cap layer is implemented yet though.

It= decodes my ~20 l2cap packets correctly at least.

> In particular I am looking for L2CAP captures containing command
&g= t; packets with InformationRequest/Response, and connectionless reception> channel packets.

These are not really in use at all at the mo= ment. I need to write a test
program to show them to you.

if they are not u= sed much, then dont waste time on them. if something is b= roken in the dissection someone will complain later.

> I would also like some captures where the ACL layer contains L2CAP
= > pdus that are fragmented across several ACL packets and a<= br>> description on how the fragmentation works for bt.

Let me se= e if I can get you some nice BNEP transmission. Once you get
BNEP on top of L2CAP you can hook it into TCP/IP then.

That would be really great. Ill start working on refactorin= g the BNEP dissector then.
Do you have a link to the standard for BNEP?<= br>

Are there also layers that map SCSI ontop of bluetooth?&nbs= p; If so it would be interesting to hook in the SCSI dissection in et= hereal into
the stack.

Regards

Marcel