On Sun, 2005-24-07 at 20:16 -0700, David S. Miller wrote:
> From: Jamal Hadi Salim <[EMAIL PROTECTED]>
> Date: Sun, 24 Jul 2005 23:11:22 -0400
>
> > ->cb could be used or made to grow dynamically ?
>
> Yes, and back to the whole SKB extensions idea and how tricky
> that is to get right and efficient at the same time.
>
indeed.
An idea we have tossed, but never really beat is of a TLVs as cb
entities - would be flexible but waste more space and consume more CPU;
> It is also a good time to remember what I said about considering
> making these sk_buff things data-local if possible. For cases that
> could fit that restriction, the only downside are object lifetime
> issues since SKB clones live a shorter amount of time than the
> associated data which can sit around forever.
we could kill about every metadata on the skb if we used a TLV maybe
with flags. The flags could indicate
things like scope {local,global}, "delete-on-copy" (which removes the
TLV), "delete-on-clone" etc. Anything along the path could pop or push
these metadata on the skb.
So something like T being 12 bits, flags taking 8 bits and L taking 12
bits.
cheers,
jamal
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
