How about this one (FT_IPv4 ignoring little_endian = 1)?
Does anyone want to take a stand?
If no arguments shoots a hole in this, how do I add the suggested change
in proto.c to CVS (who approves that)?
/ Peter
Peter Johansson wrote:
The protocol I am writing a dissector for sends (among other things)
the IP address that the data connection should be made on (similar to
FTP).
The problem I have is that the IP v4 address is in little endian
format in the protocol data. Hence I tried:
proto_tree_add_item(dsd_tree,
hf_dsd_ip_address_1,
tvb,
OCP_HEADER_SIZE + DSD_HEADER_SIZE + 6,
4,
TRUE); /* DATA IS LITTLE ENDIAN */
hf_dsd_ip_address_1 is defined as:
&hf_dsd_ip_address_1,
{
"IP address 1", "ocp.dsd.addr_1",
FT_IPv4, BASE_NONE, NULL, 0x0,
""
}
The problem is that the address that should be shown as i.e.
192.168.1.2 is shown as 2.1.168.192 since proto_tree_add_item in
proto.c ignores the little_endian flag for the type FT_IPv4.
Is this by design or is it a misfeature (bug)?
I suggest a change in proto.c from:
proto_tree_set_ipv4(new_fi, value);
to:
proto_tree_set_ipv4(new_fi, little_endian ? GUINT32_TO_BE(value) :
value);
Since proto_tree_set_ipv4 expects big endian data, the value must be
converted to BE if LE.
diff -u information follows:
--- proto.c Fri Aug 13 00:41:54 2004
+++ modified_proto.c Thu Sep 02 12:06:25 2004
@@ -722,7 +722,7 @@
case FT_IPv4:
g_assert(length == 4);
tvb_memcpy(tvb, (guint8 *)&value, start, 4);
- proto_tree_set_ipv4(new_fi, value);
+ proto_tree_set_ipv4(new_fi, little_endian ?
GUINT32_TO_BE(value) : value);
break;
case FT_IPXNET:
/ Regards, Peter
_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev
--
____________________________________________________________
Peter Johansson Phone : +46 (0)8 752 81 00
Contactor Data AB Mobile: +46 (0)70 608 3063
Finlandsgatan 14 MailTo:Peter.Johansson@xxxxxxxxxxxx
164 74 Kista SMS :Peter.Johansson.sms@xxxxxxxxxxxx
SWEDEN