Re: H.225/ASN1 problem

Kukosa, Tomas wrote:
> You are right, I have found it too.
>  
> It will be fixed ASAP.

10x

while we are at it I found another issue with val_start in packet-ber


line 1528:


if (is_per_byte_aligned){
	BYTE_ALIGN_OFFSET(offset);
	offset+=length*8;
	val_start = offset>>3;

this leaves val_start pointing to the next element and breaks things badly.

If I exchange the 2 lines, the fast start dissector works fine again ;)


the patch looks like this:


@@ -1526,8 +1527,8 @@
                         /* align to byte */
                         if (is_per_byte_aligned){
                                 BYTE_ALIGN_OFFSET(offset);
+                               val_start = offset>>3;
                                 offset+=length*8;
-                               val_start = offset>>3;
                         }else{
                                 guint8 *buff;
                                 guint32 i = 0;



Best regards,

Cvetan

>  
> 
> ________________________________
> 
> Od: [email protected] za uživatele Cvetan Ivanov
> Odesláno: pá 2.6.2006 17:32
> Komu: Ethereal development
> Předmět: Re: [Ethereal-dev] H.225/ASN1 problem
> 
> 
> 
> Hello,
> 
> The problem seems to be in the per decoder, probably
> dissect_per_octet_string
> 
> in the case
>          } else if ((min_len==max_len)&&(min_len<65536)) {  /* 16.7 if
> length is fixed and less than to 64k*/
> ...
>                  if (is_per_byte_aligned){
>                          BYTE_ALIGN_OFFSET(offset);
> 
> val_start is not updated, and if the offset was not aligned from the
> start, val_start points to the previous byte
> 
>                  if (is_per_byte_aligned){
>                          BYTE_ALIGN_OFFSET(offset);
> +                       val_start = offset>>3;
>                          val_length = min_len;
>                          offset+=min_len*8;
> 
> seems to fix it, but it would be better if somewone with deeper knolwdge
> about packet-ber takes a look at it
> 
> btw it would be nice to replace "per_aligment_type_callback" with
> "per_alignment_type_callback" in the sources
> 
> 
> Cvetan Ivanov wrote:
>> Hello,
>>
>> While trying few captures with the current svn, I noticed that ip
>> addresses and GUIDs are decoded incorrectly.
>>
>> example:
>>
>> we have source signall address element that looks like this:
>>
>> 00 01 02 03 04 05 06
>>
>> where 0 means ipAddress and the IP should be 1.2.3.4 port 0x0506
>> but it's decoded instead as 0.1.2.3 port 0x0506
>>
>> the same thing goes for the CallIdentifier -
>>
>> 00 01 02 .... (17 bytes) is decoded like guid 000102..., and it should
>> be 010203...
>>
>> Tested on two different linux systems. Unfortunatrly I'm not too much
>> into asn1 and don't have idea how to debug this further.
> 
> 
> best regards,
> 
> Cvetan
> 
> 
> --
> Cvetan Ivanov
> System Administrator
> SpectrumNet Jsc.
> +35929657613 Office
> _______________________________________________
> Ethereal-dev mailing list
> [email protected]
> http://www.ethereal.com/mailman/listinfo/ethereal-dev
> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Ethereal-dev mailing list
> [email protected]
> http://www.ethereal.com/mailman/listinfo/ethereal-dev

-- 
Cvetan Ivanov
System Administrator
SpectrumNet Jsc.
+35929657613 Office