Huge thanks to our Platinum Members Endace and LiveAction,
and our Silver Member Veeam, for supporting the Wireshark Foundation and project.

Ethereal-users: Re: [Ethereal-users] RE:capture using two serial ports for serial ppp

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: Guy Harris <gharris@xxxxxxxxx>
Date: Fri, 31 Dec 2004 12:11:37 -0800
Frank v. Münchow-Pohl wrote:

when my understanding is correct, the requested feature
should be provided in packet.dll (when using winpcap as capture
interface), not in ethereal itself.

No, it should be in libpcap, so that it can be provided when WinPcap *isn't* being used as a capture interface - WinPcap doesn't work very well on my primary laptop, for example, except perhaps in a Virtual PC for Mac window, and won't work on my other laptop except perhaps in Windows-under-VMWare on a Linux or perhaps FreeBSD partition.

There'd probably still have to be some differences between the Win32 and UN*X versions of the code, as the calls to manipulate serial ports would be different, but a lot of the code (e.g., code to extract frames and remove escapes) could be shared by the Win32 and UN*X versions.

I realized, that handling VJ- and PPP-compression
would be a lot of work, but I don't need them in my work now.
It seems to me, that these features are handled in ethereal
rather than in winpcap, but to make ethereal understand compressed
data through a pcap file, they have to be decoded (Or is there a way
to tell ethereal to decode these packets?).

As long as information as to which direction the packets were travelling is available, Ethereal can decode that automatically, at least for VJ compression, as there's a PPP protocol type for that, and Ethereal recognizes it.

Unfortunately, the current libpcap format doesn't have that information. You could try having your stand-alone program write the file in pppdump format, which does have that information - and which contains raw bytes from the serial line, so the code wouldn't have to extract frames and remove escape sequences (although if it works in that fashion it wouldn't fit into libpcap/WinPcap, which expect frames to be delivered).