Fix an obvious error in the nfs4 stateid parsing. The stateid is used in a number of common operations (such as open and setattr), so this caused a lot of misparsing.
svn path=/trunk/; revision=20700
* fields of an uat table now are passed using an array of uat_filed_t
* field callbacks take two more userdata arguments
* add some macros to define uat field callbacks.
* uats can be registered as preferences for a specific protocol
- the preference widget is a button that opens the uat's window
* dfilter-macro => reflect changes to API
svn path=/trunk/; revision=20695
Wed, Jan 31, 2007 at 7:24 PM
To: wireshark-dev@wireshark.org
Hello,
Please consider for checkin the following new dissectors, for the FMP protocol.
FMP (File Mapping Protocol) is the network protocol basis for EMC's HighRoad (MPFS) technology. Highroad is used to allow multiple clients to share access to NAS-shared files while allowing clients to directly access data volumes (via, for example, Fibre Channel or iSCSI). EMC currently uses this technology in our Celerra NAS servers, and we're currently in the process of open sourcing portions of the technology.
FMP actually consists of two ONC/RPC-based protocols - the core FMP protocol, and FMP/Notify. The latter is used as an asynchronous callback to inform clients of status changes, such as lock revocation.
We'd like to offer these dissectors to Wireshark users for help in debugging or otherwise troubleshooting MPFS-related problems. There are still a few minor changes that need to be made ( i.e. a handful of fields that aren't decoded) but the dissector is overall fairly complete and very usable.
Let me know if there are questions or feedback, or otherwise if other info is needed (like sample captures, which I don't want to send out to the mailing list).
Thanks,
Ian Schorr
EMC Corporation
svn path=/trunk/; revision=20679
the changes in r20642 to release-notes.xml. Add "news" targes to the
docbook makefiles to copy release-notes.txt to NEWS.
svn path=/trunk/; revision=20669
in addition to the currently supported two or one digit at a time. This
applies in particular to the FT_ETHER MAC addresses, so that a MAC address
can now be entered as xxxx.xxxx.xxxx in addition to the traditional formats.
svn path=/trunk/; revision=20660
IMO, the function should return an error string instead of cowardly
bailing out with an empty string.
inet_ntop does the latter by default, so it does not need any additional
check, just passing buf_len instead of INET6_ADDRSTRLEN.
At the other side I like the check you made into ip_to_str_buf.
My proposal would be to combine the patches and use MAX_IP_STR_LEN,
and approach below for both IPv4 and IPv6 in address_to_str_buf.
Have the error string return a more descriptive clue.
svn path=/trunk/; revision=20658
- most paramaters have data set to NULL, have them added to the tree
- avoid the context tracing mecanism bailing out if a ContextList appears in the packet.
- in q9150 do not attempt to dissect sdp unless we believe it is sdp
(BTW we need heuristics for sdp and other potential payloads of this parameter..)
svn path=/trunk/; revision=20649