After a very quick review here are some changes:
1, you have to terminate all value_strings with a
{0, NULL}
entry to prevent them from reading beyond the end of the array.
2, in dissect_pvfs2_error()
remove the if (err == 0) and always add the item using
proto_tree_add_uint() regardless of what error is. Add a { 0,
"Success"}, to the value_string for error.
3, in dissect_pvfs2_error() add code so that IF there was a err!=0
then do the COL_INFO stuff to add the error to the information part of
the summary line
4, dissect_pvfs2_attrmask()
Should you not have this one dissect a lot of proto_tree_add_boolean()
with FT_BOOLEANs ?
5, can you replace all strcpy() (and strcat() if you use them) with
g_snprintf()
(i did not check for whether there could be infinite loops where
offset is never updated)
can you do these changes and resubmit for another review?
can you also add a wiki page for the protocol and upload an example
capture to the wiki.
On 10/7/05, Mike Frisch <mfrisch@xxxxxxxxxx> wrote:
> Enclosed is the source for the dissector for the Parallel Virtual File
> System (PVFS). The dissector specifically implements support for PVFS
> version 2, however it is left open to support PVFS version 1.
>
> More about PVFS can be found at http://www.pvfs.org.
>
> I also have several sample captures at
> http://dvbwave.com/pvfs2-captures.tar.bz2. These sample captures were
> used for fuzz testing as well.
>
>
>