Wireshark · Ethereal-dev: RE: [Ethereal-dev] Wrong decoding of PER Enumerated type withextension?

["Anders Broman (AL/EAB)" <anders.broman@xxxxxxxxxxxx>]

Hi,
Great, I think we should do it according to your proposal and do it now.

Brg
Anders

Minor change:
/* 13 Enemerated */
guint32
dissect_per_enumerated(tvbuff_t *tvb, guint32 offset, packet_info
*pinfo, proto_tree *tree, int hf_index, guint32 min, guint32 max,
guint32 *value, proto_item **item, gboolean has_extension, guint32
ext_min, guint32 ext_max)
{

	proto_item *it=NULL;
	guint32 val;
	proto_item *pi;
	guint32 start_offset = offset;
	gboolean extension_present;
	header_field_info *hfi;

	if (!has_extension){
		/* 13.2  */
		offset = dissect_per_constrained_integer(tvb, offset,
pinfo, tree, hf_index, min, max, value, item, has_extension);
		return offset;
	}
	/* Just get the extension bit, don't advance offset as it will
be done in
	 * dissect_per_constrained_integer
	 */
	dissect_per_boolean(tvb, offset, pinfo, tree, -1,
&extension_present, &pi);
	if(extension_present){
		/* Add extension bit to the tree */
		offset = dissect_per_boolean(tvb, offset, pinfo, tree,
-1, &extension_present, &pi);
		if (!display_internal_per_fields)
PROTO_ITEM_SET_HIDDEN(pi);

		hfi = proto_registrar_get_nth(hf_index);
		/* 13.3  */
		if ( ext_min == ext_max ){
		
			/* 10.5.4	If "range" has the value 1,
			 * then the result of the encoding shall be
			 * an empty bit-field (no bits).
			 */
			val = max + 1;
			if (IS_FT_UINT(hfi->type)) {
				it = proto_tree_add_uint(tree, hf_index,
tvb, start_offset, 1, val);
			} else {
				THROW(ReportedBoundsError);
			}

			/* byte aligned */
			BYTE_ALIGN_OFFSET(offset);
			if (item) *item = it;
			if (value) *value = val;
			return offset;

		}else{
			/* 13.3 ".. and the value shall be added to the
field-list as a
			 * normally small non-negative whole number
whose value is the 
			 * enumeration index of the additional
enumeration and with "lb" set to 0.."
			 *
			 * Byte align after reading Extension bit 
			 *
			 */
			BYTE_ALIGN_OFFSET(offset);
			offset=dissect_per_integer(tvb, offset, pinfo,
tree,
				hf_index, &val, item);
			val = val + max;
			if (IS_FT_UINT(hfi->type)) {
				it = proto_tree_add_uint(tree, hf_index,
tvb, start_offset, 1, val);
			} else {
				THROW(ReportedBoundsError);
			}
			if (item) *item = it;
			if (value) *value = val;
			return offset;
		}

	}
	/* Extension not present */

	offset = dissect_per_constrained_integer(tvb, offset, pinfo,
tree, hf_index, min, max, value, item, has_extension);

	return offset;
}

-----Original Message-----
From: ethereal-dev-bounces@xxxxxxxxxxxx
[mailto:ethereal-dev-bounces@xxxxxxxxxxxx] On Behalf Of Kukosa, Tomas
Sent: den 6 mars 2006 11:37
To: Ethereal development
Subject: RE: [Ethereal-dev] Wrong decoding of PER Enumerated type
withextension?

Hi Anders,

I will update asn2eth in accordance with your changes in packet-per.

But I guess we should change dissect_per_enumerated() /now or later/ It
does not support ENUMARATED where values differs from indexs Cause ::=
ENUMERATED {
   cause0,
   cause20 (20),
   cause30 (30),
   ...
   cause100 (100)
}

My proposal is following
- remove 'guint32 min' parameter as it is always 0
- rename 'guint32 max' to 'root_num' and value will be number of root
items
- remove 'guint32 ext_min' parameter as it is always 0
- rename 'guint32 ext_max' to 'ext_num' and value will be number of
extension items
- add new parameter 'guint32 *value_map' which maps indexe to values /if
NULL root vaules matches indexes and extension values matches
root_num+index/

for upper example it would be: 
root_num = 3
ext_num = 1
value_map = {0, 20, 30, 100}

Any commets?

Regards,
  Tomas




Mailcode: V55BCPWP
-----Original Message-----
From: ethereal-dev-bounces@xxxxxxxxxxxx
[mailto:ethereal-dev-bounces@xxxxxxxxxxxx] On Behalf Of Anders Broman
Sent: Saturday, March 04, 2006 5:49 PM
To: ethereal-dev@xxxxxxxxxxxx
Subject: [Ethereal-dev] Wrong decoding of PER Enumerated type with
extension?

Hi,
I have a (k12)RANAP trace which displays as malformed the affected ASN.1
code is:

asn1/ranap/ranap.asn:
PagingCause ::= ENUMERATED {
        terminating-conversational-call,
        terminating-streaming-call,
        terminating-interactive-call,
        terminating-background-call,
        terminating-low-priority-signalling,
        ...,
        terminating-high-priority-signalling
}
The trace is encoding "terminating-high-priority-signalling" as 0x80
reading
X.691 section 13:
13.1    The enumerations in the enumeration root shall be sorted into
ascending order by their enumeration value, and shall then be assigned
an enumeration index starting with zero for the first enumeration, one
for the second, and so on up to the last enumeration in the sorted list.
The extension additions (which are always defined in ascending order)
shall be assigned an enumeration index starting with zero for the first
enumeration, one for the second, and so on up to the last enumeration in
the extension additions.

NOTE - ITU-T Rec. X.680 | ISO/IEC 8824-1 requires that each successive
extension addition shall have a greater enumeration value than the last.

13.2    If the extension marker is absent in the definition of the
enumerated type, then the enumeration index shall be encoded. Its
encoding shall be as though it were a value of a constrained integer
type for which there is no extension marker present, where the lower
bound is 0 and the upper bound is the largest enumeration index
associated with the type, completing this procedure.

13.3    If the extension marker is present, then a single bit shall be
added
to the field-list in a bit-field of length one. The bit shall be set to
1 if
the value to be encoded is not within the extension root, and zero
otherwise. In the former case, the enumeration additions shall be sorted
according to 13.1 and the value shall be added to the field-list as a
normally small non-negative whole number whose value is the enumeration
index of the additional enumeration and with "lb" set to 0, completing
this procedure. In the latter case, the value shall be encoded as if the
extension marker is not present, as specified in 13.2.

I guess the encodings should be " Extension bit set and value = 0, range
of value is 1 so according to:

10.5.4	If "range" has the value 1, then the result of the encoding
shall be
an empty bit-field (no bits).

This looks OK(?)

To decode this properly I suppose a new dissect_per_enumerated() has to
be made and asn2eth needs to supply both the range for the enumeration
root and the extension to this funktion?

Best regards
Anders


_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev

_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev