Re: [mpls-linux-general] fec binding

[James R. L. <jl...@mi...>]

*sigh*  You are correct.  When I origianlly implemented mpls-linux the reason
I stayed away from iptables was exactly this reason.  For some reason when I
was cosidering the move to iptables again, I completly ignored my "this doesn't
feel right" sensation.  I'm now regretting my change to iptables.

... but it is such a cleaner implementation then before ...

I will see if there is any way to get iptable (netfilter) to do what I want
(exact match in the FIB) otherwise I will have to add the FIB based FTN
again.

*sigh*

Thank you for being the voice of reason ....

Jim

On Fri, Feb 08, 2002 at 01:44:54PM +0100, Yon Uriarte wrote:
> Hello,
> 
>  I've just catched up some months list life. I see with concern that it
> is no longer possible to bind a routing table entry to a label.
> 
>  Isn't the semantic of vanilla (ip routing table driven) ldp the longest
> match selection? The FTN mib matches the iptables format almost perfectly
> otherwise. But for longest match selection it is not very adequate. For
> example, given the routing table
> 1.1.1.0/24
> 1.2.0.0/16
> 1.0.0.0/8
> and a ldp mapping for 1.0.0.0/8 at least 3 iptables entries have to be
> created:
> -d 1.1.1.0/24 -j RETURN
> -d 1.2.0.0/16 -j RETURN
> -d 1.0.0.0/8 -j MPLS
> 
>  And I guess implementing this fec -> iptables behavior in the guts of the
> ports will be kind of a pain.
> 
>  Or are my assumptions about semantics wrong?
> 
> TIA, HAND
>  yon
> 
> 
> 
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu