On Thu, 2009-07-09 at 19:23 -0700, Peter Memishian wrote:
> > >  > http://cr.opensolaris.org/~seb/clearview-libpcap/
>  > >  > 
>  > >  > The idea is to translate DL_IPV4, DL_IPV6, and DL_6TO4 to DLT_RAW, 
> which
>  > >  > indicates that captured packets begin with an IP header.  Wireshark 
> can
>  > >  > then decode them seamlessly without any modification to the 
> application.
>  > > 
>  > > The use of DL_6TO4 to gate the tunnel stuff seems to be a bit chummy with
>  > > the history of Solaris tunnels; how would a future libpcap developer
>  > > understand why that's correct?
>  > 
>  > Ah, you mean it would be clearer if we #ifdef'ed each DL_* value
>  > individually?  We could certainly do that.  I used DL_6TO4 simply
>  > because I knew that if it exists, the other two exist as well...
> 
> It seemed to go a step further than that and make the assertion that
> DL_6TO4 indicated support for Clearview IP Tunneling.  Otherwise, it
> would make sense to also enable the support if e.g. DL_IPV4 was defined
> but DL_6TO4 was not.

Yeah, that's fine.  I've updated the webrev.  On a related note, I guess
SFW build machines will need to be running post-clearview bits to get
the libpcap support compiled in.  I wonder how they usually synchronize
features like this on build machines...

-Seb



Reply via email to