Re: [ATM] who 'owns' the skb created by drivers/atm?

[Mitchell Blank Jr]

David S. Miller wrote:
> On Thu, 2003-02-20 at 17:42, James Morris wrote:
> > skb->cb is owned by whatever layer is currently processing the skb.
> 
> Furthermore, once you netif_rx() an SKB it is no longer yours.
> It is owned by the networking stack.
> 
> If the ATM layer wants to do fancy things and still pass the SKB
> to netif_rx(), _it_ should clone the SKB and give that clone to
> the ATM layer directly.

As far as I'm aware the ATM layer doesn't care what happens to the
SKB after it gets passed to netif_rx(), so I don't know why this would
be a problem.  Some people seem to be suggesting that we need to zero
out ->cb before passing the SKB to netif_rx() but I don't see why
that would be neccesary.

-Mitch