> >This is not a issue with the kernels after 2.1.73 or so as 0.0.0.0 is not
> >a valid address for network interfaces..
>
> Switching to 2.1.x is not an immediate option for me, so the 2.0.x code
> ought to be fixed.
I would suggest testing the fix and submitting it for 2.0.34..
>
> Why is 0.0.0.0 not a valid address for network interfaces, at least during
> DHCP initialization?
See standards, or alan cox, *shrugs*.
>
> >Tho it should be noted that how the 2.0 DHCP client's work is actualy
> >broken according to the standards..
>
> You mean by its use of 0.0.0.0? No, this is what is required by
> RFC2131. See page 22 where it specifically says that client messages
> must have a source address of 0.0.0.0 until it receives an IP address
> assignment. Also, there needs to be some way to receive the DHCPOFFER
> response from the server -- which will be to an as-yet-unknown IP
> destination address.
You would have to take this one up with Alan Cox, but I think its allready
been hashed over..
>
> I guess one way to do DHCP without hacking the IP interface code is to
> implement it on top of a raw Ethernet socket, if such a thing is even
> implemented in Linux. But it seems a rather drastic change when a much
> simpler and cleaner short-term fix is available at least for the 2.0.x
> kernels.
Thats what has been deamed to be the correct way to handle things..
>
Zephaniah E, Hull
> Phil
>
>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu