BPF on Solaris?

[David W. Hankins]

On Tue, Sep 13, 2005 at 02:11:38PM -0700, Ted Lemon wrote:
> On Sep 13, 2005, at 12:57 PM, David W. Hankins wrote:
> > I was under the impression that wasn't a reliable mark of wether
> > or not the client broadcast the packet to the segment from which
> > it was received.
> 
> If the client sends a broadcast, the destination address you receive  
> should be 255.255.255.255.   So you don't know what subnet the client  

Should.  Mightnt it also be addressed to 192.168.0.255, even if the
subnet from whence the packet was received was not rfc1918 addressed?

Should that be treated like a unicast renewal?

I'm really uncomfortable with the odds.

I think to do this properly you need both the ip and mac layer
destinations.  If either is broadcast, treat it like one.

> is on, but you do know that it sent a broadcast.   You can determine  
> to which network the client is attached by looking at the interface  
> ID, which I think you can get from Solaris' socket API with a recvmsg 
> () option.

Yes, I think most of the pieces might be there, and this is stuff
that should be done anyway for platforms that don't have a bpf or
dlpi or etc.  It would help substantially.

-- 
David W. Hankins		"If you don't do it right the first time,
Software Engineer			you'll just have to do it again."
Internet Systems Consortium, Inc.		-- Jack T. Hankins