Wireshark-bugs: [Wireshark-bugs] [Bug 8691] Adding support of BGP flow spec RFC 5575
Date: Sun, 19 May 2013 16:52:49 +0000

Comment # 9 on bug 8691 from
Hi Matt,

Some example :
+        proto_tree_add_boolean(op_value_tree,
hf_bgp_flowspec_nlri_tcp_flags_cwr, tvb, offset+cursor_op_val, 1, tcpflags);
+        proto_tree_add_boolean(op_value_tree,
hf_bgp_flowspec_nlri_tcp_flags_ecn, tvb, offset+cursor_op_val, 1, tcpflags);
+        proto_tree_add_boolean(op_value_tree,
hf_bgp_flowspec_nlri_tcp_flags_urg, tvb, offset+cursor_op_val, 1, tcpflags);
+        proto_tree_add_boolean(op_value_tree,
hf_bgp_flowspec_nlri_tcp_flags_ack, tvb, offset+cursor_op_val, 1, tcpflags);

Must be replace by proto_tree_add_item

Also this code :
+        proto_tree_add_boolean(op_value_tree, hf_bgp_flowspec_nlri_op_eol,
tvb,offset+cursor_op_val, 1,operator);
+        proto_tree_add_boolean(op_value_tree, hf_bgp_flowspec_nlri_op_and,
tvb,offset+cursor_op_val, 1,operator);
+        proto_tree_add_text(op_value_tree, tvb, offset+cursor_op_val, 1,
"Value length: %u (1 << %u)",value_len,shift_amount);
+        proto_tree_add_boolean(op_value_tree, hf_bgp_flowspec_nlri_op_flg_not,
tvb,offset+cursor_op_val, 1,operator);
+        proto_tree_add_boolean(op_value_tree,
hf_bgp_flowspec_nlri_op_flg_match, tvb,offset+cursor_op_val, 1,operator);

is duplicate.
It is not possible to make a function to decode the op ? and return the length
(and also replace by proto_tree_add_item)

+        if (value_len == 1) {
+        proto_tree_add_text(op_value_tree, tvb, offset+cursor_op_val,
+                            value_len, "%s : %u",tag,tvb_get_guint8(tvb,
offset+cursor_op_val));
+        }
+        else {
+        proto_tree_add_text(op_value_tree, tvb, offset+cursor_op_val,
+                                value_len, "%s : %u",tag,tvb_get_ntohs(tvb,
offset+cursor_op_val));
+        }
you should use hf with FT_UINTXX and specify the length in proto_tree_add_item
function (4eme arg)

+        proto_tree_add_text(op_value_tree, tvb, offset+cursor_op_val,
+                            value_len, "%s : 0x%02x
%s",tag,dscp_value,val_to_str_ext_const(dscp_value,&dscp_vals_ext, "Unknown
DSCP"));
+        /*proto_tree_add_item(op_value_tree, hf_bgp_flowspec_nlri_dscp, tvb,
offset+cursor_op_val, 1, ENC_NA);*/
there is a problem with hf (proto_tree_add_item function) ? 


+            proto_tree_add_text(nlri_tree, tvb, offset+cursor_fspec,
+                                1,"Type destination prefix filter");
[...]
+            proto_tree_add_text(nlri_tree, tvb, offset+cursor_fspec,
+                                    1,"Type source prefix filter");

Look the dev guide about value_string (for create a table with all type)


You are receiving this mail because:
  • You are watching all bug changes.