On Sun, Jan 30, 2011 at 1:15 PM, Yosi Saggi <[email protected]>wrote:

> Hi Martin
>
>
>
> I am going over the mac_lte_info struct, and wanted to know if you can help
> me with the msain parameters I see I need to input to decode.
>
> Most of the parameters I understand and can fill in the 2 of the top 3 I
> don’t understand:
>
>    guint8          radioType;
>
>    guint8          rntiType;
>
>
Your follow-up email said you now understand how to fill these fields in.


>
>
> Direction is obvious.
>
> What does the “radioType” stand for? I looked for some reference in the
> standard, but I guess I just don’t know what to look for.
>
> What does “rntiType” stands for? Is it referring to SI-RNTI, RA-RNTI and
> C-RNTI?
>
>
>
> I am also not sure about the isPredefinedData, reTxcount, dl_retx and
> oob_event parameters.
>
>
>

I have just committed a change to packet-mac-lte.h that hopefully gives
clear descriptions of how these fields work.
Basically if the struct is zeroed (as se_alloc0() does) and you only fill in
the parts you want to the rest should be ignored.
See http://anonsvn.wireshark.org/viewvc?view=rev&revision=35715 for my
changes.  Please note that the 1.4 branch is quite far from being up to date
with the LTE protocols so the trunk is much more up-to-date.


> Is there a capture sample for the LTE that I can see how the dissector
> works?
>
>
>
I will try to dig out a simple test file for you tomorrow.


> I do not use catapult-dct-2000. Also the LTE packets I am capturing are not
> over UDP.
>
> My dissector is between PHY and MAC. I get Ethernet packets with type FFFF
> and dissect them with our proprietary API. I have no problem with my own
> API, the problem is when I’m getting to the MAC level (to the PHY data
> level). This is why I want to use the MAC, RLC,PDCP dissectors. I just don’t
> understand what do those parameters in the struct mean so I could fill them
> up.
>
>
>

I'm imagining that your own dissector is registered with ethertype FFFF, and
then you then parse some parameters from the start of each frame before you
come to the MAC payload.  And that those parameters include enough of
mac_lte_info to get useful decodes.  Do let me know if any fields are still
unclear.  Since my company's (test) product has been the first user, the
fields available are probably a bit biased towards the UE side of the link
and the way we happen to log them.  But there are now other users of the UDP
framing method and the more people using the dissectors (and hopefully
helping to improve them) the better (and more generally useful) they'll be.


> Thanks for the help
>
>
>
> Yosi
>
>
>

Best regards,
Martin


>
>
> *From:* [email protected] [mailto:
> [email protected]] *On Behalf Of *Martin Mathieson
> *Sent:* Thursday, January 27, 2011 6:31 PM
>
> *To:* Developer support list for Wireshark
> *Subject:* Re: [Wireshark-dev] wireshark crashes after adding
> preferencecodeto my
>
>
>
> Hi Yosi
>
> On Thu, Jan 27, 2011 at 3:37 PM, Yosi Saggi <[email protected]>
> wrote:
>
> When I call the “mac-lte” dissector I get this message:
>
> "Can't dissect LTE MAC frame because no per-frame info was attached!"
>
>
>
> You need to create a struct of type mac_lte_info (see packet-mac-lte.h),
> fill in at least some of the details, and attach it to the frame using
> p_add_proto_data.
>
> There are 2 examples of this:
> - attach_mac_lte_info() in packet-catapult-dct2000.c (which I guess you're
> not using)
> - dissect_mac_lte_heur() in packet-mac-lte.c.  This is enabled by turning
> on a preference "Try heurisitc LTE-MAC over UDP framing".  See the wiki page
> for a link to an example program that sends these frames
>
>
> I don’t know what info do I need to add to the “pinfo” so it would work.
>
> If you can help me in this point I will be very grateful.
>
>
>
> I tried to send to the “mac-lte” dissector the buffer that contains only
> the MAC data (including the headers of course and all that comes with it
> (RLC, PDCP and finally IP data).
>
>
>
>
> The dissector needs the context that comes from the mac_lte_info struct to
> decode the frame properly.  Some of the parameters are optional but useful
> for troubleshooting.  By setting preferences appropriately you can decode
> all the way up to RRC or IP, and there are stats (from the Telephony menu)
> for MAC and RLC which I use almost every day.
>
> If you have the fields that go into mac_lte_info in a different format
> (from the existing heuristic header), you could write a simple dissector
> that just parsed out the header, filled in the struct and called the mac_lte
> dissector with the payload.
>
> Best regards,
> Martin
>
> Any help will be great.
>
>
>
> Yosi
>
>
>
> *From:* [email protected] [mailto:
> [email protected]] *On Behalf Of *Martin Mathieson
> *Sent:* Wednesday, January 26, 2011 12:44 PM
>
>
> *To:* Developer support list for Wireshark
> *Subject:* Re: [Wireshark-dev] wireshark crashes after adding preference
> codeto my
>
>
>
> I can't see anything immediately wrong with this.  I assume your prefs
> variables (global_dan_lte_sdk_dissect_MAC and
> global_dan_lte_sdk_dissect_RLC) are gbooleans with file scope?
>
>
>
> I'm wondering why the existing LTE MAC dissector isn't suitable, or can't
> be made to work with your frames?
>
> Regards,
> Martin
>
> On Wed, Jan 26, 2011 at 10:07 AM, Yosi Saggi <[email protected]>
> wrote:
>
> Hi
>
> I wanted to add some options in the preference window for my dissector.
> This is the code I have added to the “proto_register” function:
>
>
>
> module_t *dan_lte_sdk_module;
>
>
>
> proto_dan_lte_sdk = proto_register_protocol ("DAN LTE SDK Protocol", "Dan
> LTE SDK", "dan_lte_sdk");
>
>
>
> register_init_routine(dan_defragment_init);
>
>
>
> proto_register_field_array (proto_dan_lte_sdk, hf, array_length (hf));
>
> proto_register_subtree_array (ett, array_length (ett));
>
> register_dissector("dan_lte_sdk", dissect_dan_lte_sdk, proto_dan_lte_sdk);
>
>
>
> dan_lte_sdk_module = prefs_register_protocol(proto_dan_lte_sdk, NULL);
>
>
>
> prefs_register_bool_preference(dan_lte_sdk_module, "Dissect_MAC_Payload",
>
>                                "Dissect MAC Layer from Data Payload",
>
>                                "In Uplink and Downlink data packets,
> dissect MAC heaser layer "
>
>                                "Disabling MAC dissection will disable RLC
> dissection automaticly",
>
>                                &global_dan_lte_sdk_dissect_MAC);
>
>
>
> prefs_register_bool_preference(dan_lte_sdk_module, "Dissect_RLC_Payload",
>
>                                "Dissect RLC Layer from Data Payload",
>
>                                "In Uplink and Downlink data packets,
> dissect RLC heaser layer ",
>
>                                &global_dan_lte_sdk_dissect_RLC);
>
>
>
> It works fine until I add those two (even one of them)
>  "prefs_register_bool_preference" functions. What did I do wrong?
>
> I get a "Runtime Error!" for wireshark.exe Removing those function fixes
>  it back.
>
>
>
> Thanks
>
>       Yosi
>
>
> ___________________________________________________________________________
> Sent via:    Wireshark-dev mailing list <[email protected]>
> Archives:    http://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
>             mailto:[email protected]?subject=unsubscribe
>
>
>
>
> ___________________________________________________________________________
> Sent via:    Wireshark-dev mailing list <[email protected]>
> Archives:    http://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
>             mailto:[email protected]?subject=unsubscribe
>
>
>
> ___________________________________________________________________________
> Sent via:    Wireshark-dev mailing list <[email protected]>
> Archives:    http://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
>             mailto:[email protected]?subject=unsubscribe
>
___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

Reply via email to