FW: binpac related question

Hi All,

Tying to use binpac to create a new protocol analyzer I run into a small issue relative to byte alignment. I am thinking this must be a known aspect however I didn’t find much documentation about it.

Padding in binpac works well inside the packet. However, sometime the packet ends with a structure that requires padding but the packet doesn’t carry the extra bytes for the padding. As a result binpac fires an exception and the parsing is terminated before extracting the data.

Here is a sample code:
type ByteArray = record{
len : uint32;
data: uint8[len];
pad : padding align 4;
};

When the last 6 bytes of a packet that ends with the above record is: 00 00 00 02 AA BB (with: len=00 00 00 02; data ={AA, BB}) binpac throws:

0.000000 binpac exception: binpac exception: out_of_bound: ByteArray: 8 > 6

Removing the padding entry from the code above results in a correct execution, however the padding is needed for the case when ByteArray is strictly inside the packet.

The solution I am thinking is way too complex (pass some arguments that will help decide if the padding is needed), but I was wondering if there is a much more elegant approach.

Another option will be to change the binpac compiler code (pac_record.cc) in method: void RecordPaddingField::GenFieldEnd(Output* out_cc, Env* env, const DataPtr& field_begin) but I am thinking this case should be covered in a different way.

Any suggestions would be greatly appreciated.

Thank you,
Laurentiu Cucos

The contents of this e-mail are intended for the named addressee only. It contains information that may be confidential. Unless you are the named addressee or an authorized designee, you may not copy or use it, or disclose it to anyone else. If you received it in error please notify us immediately and then destroy it.

Hi Laurentiu,

Is the padding always at the end? If so, you can use:

padding: bytestring &restofdata;

Ruoming

2008/12/1 Cucos, Laurentiu <Laurentiu.Cucos@compuware.com>

Hi Ruoming,

Thanks for the answer. First, I would like to say how impressed I am with binpac capabilities and the approach used. Once I had a test environment setup and went over the samples and documentation, I was able to develop a fairly extensive decoder for a relatively complex protocol (IIOP) in just a day or so (took longer to go over the protocol specs).

Regarding the issue in question, the padding is not always at the end. Depending on some protocol data (i.e. version), the ByteArray record can be either in mid packet or at the end.
The things are complicated by the fact that this record is part of other records (that ultimately depend on version) - and I would not like to duplicate records with every version.

As of now I found a solution by changing: pac_record.cc:

void RecordPaddingField::GenFieldEnd(Output* out_cc, Env* env, const DataPtr& field_begin)

case PAD_TO_NEXT_WORD:

// new code: disable the padding if exceeds the packet size (not sure about all implications)
out_cc->println("%s = (%s == 0 || (%s + %s > %s)) ? 0 : %d - %s;",
padding_var,
padding_var,
padding_var,
field_begin.ptr_expr(),
env->RValue(end_of_data),
wordsize_,
padding_var);

// original code:
// out_cc->println("%s = (%s == 0) ? 0 : %d - %s;",
// padding_var,
// padding_var,
// wordsize_,
// padding_var);

Any feedback will be highly appreciated.

Thank you,
Laurentiu