I agree and see that you committed it already :-)
On Tue, 24 Jul 2001, Brooks Davis wrote:
> On Tue, Jul 24, 2001 at 05:04:53PM -0700, Julian Elischer wrote:
> > On Tue, 24 Jul 2001, Brooks Davis wrote:
> >
> > > Index: ng_split.c
> > > ===================================================================
> > > RCS file: /home/ncvs/src/sys/netgraph/ng_split.c,v
> > > retrieving revision 1.1
> >
> > [...]
> >
> > > diff -u -r1.1 ng_split.c
> > > - /*
> > > - * XXX Really here we should just remove metadata we understand.
> > > - */
> > > - NGI_GET_META(item, meta);
> > > - NG_FREE_META(meta);
> >
> >
> > this one is tricky..
> > it was written as part of a set of nodes that pass around their own
> > metadata. The original idea was (so the author suggested) to stop those
> > metadata structures from propogating out of the limited part of the graph
> > that knew about them.
> >
> > However I don't see any harm in letting them go, since any node that
> > doesn't understand a particular metadata type should ignore it..
>
> My idea was that split falls into the same catagory as one2many or tee
> in that their sole function is to move packets around and thus removing
> metadata was a POLA violation. It seems that if you do need to create a
> wall to keep metadata out, the right answer would be an ng_stripmeta
> node who's purpose would be to remove metadata to provide this kind of
> isolation where needed.
>
> -- Brooks
>
> --
> Any statement of the form "X is the one, true Y" is FALSE.
> PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4
>
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message