netdev
[Top] [All Lists]

Re: [IPSEC] Move hardware headers for decaped packets

To: kuznet@xxxxxxxxxxxxx
Subject: Re: [IPSEC] Move hardware headers for decaped packets
From: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
Date: Sun, 7 Dec 2003 20:02:05 +1100
Cc: "David S. Miller" <davem@xxxxxxxxxx>, jmorris@xxxxxxxxxx, netdev@xxxxxxxxxxx
In-reply-to: <20030925124102.GA18188@gondor.apana.org.au>
References: <20030925121131.GA17968@gondor.apana.org.au> <200309251228.QAA11650@yakov.inr.ac.ru> <20030925124102.GA18188@gondor.apana.org.au>
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mutt/1.5.4i
On Thu, Sep 25, 2003 at 10:41:02PM +1000, herbert wrote:
> 
> A quick grep shows just below 100 occurances of mac.raw under
> drivers/net.  I'll have a go at it unless someone else does
> it first.
> 
> On the other hand, I've found a way to avoid the move without
> requiring mac_len to be set at all.  We can do this in netif_rx:

It looks like I've neglected this issue for a few months.

I have had another look at it and avoiding the move altogether
is actually not that useful.  The reason is that anything that
uses mac.raw will need to handle the case where the MAC header
is not next to the payload.

In particular, this is non-trivial in the AF_PACKET case so
we'll end up moving the MAC header anyway.  Worse yet, if there
are multiple AF_PACKET users, this means that the copy/move
will have to be repeated.

So I'd like to go with the previous approach which is to move
for tunnel SAs only by using mac_len.

Here is that patch again.
-- 
Debian GNU/Linux 3.0 is out! ( http://www.debian.org/ )
Email:  Herbert Xu ~{PmV>HI~} <herbert@xxxxxxxxxxxxxxxxxxx>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Attachment: p
Description: Text document

<Prev in Thread] Current Thread [Next in Thread>