From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752188AbZBQJcj (ORCPT ); Tue, 17 Feb 2009 04:32:39 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751228AbZBQJc1 (ORCPT ); Tue, 17 Feb 2009 04:32:27 -0500 Received: from mail-bw0-f161.google.com ([209.85.218.161]:55602 "EHLO mail-bw0-f161.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751131AbZBQJc0 (ORCPT ); Tue, 17 Feb 2009 04:32:26 -0500 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=date:from:to:cc:subject:message-id:references:mime-version :content-type:content-disposition:in-reply-to:user-agent; b=BWmdsNv2EzCF97+ZGJdN2G8VSuLUQq51joTsTGtldq58oX5Yf+pNJXo2Zenp831Nfg pbHSEXgOqnbW2NYLg8FliimtNYRdTMqSeXoQ+i1CBc7HOeyaVDKs1fSGi+1B3cxXHqy3 BbIKY06LO9Or4fFbZWWZcnYujFN+P5r+CCnhk= Date: Tue, 17 Feb 2009 09:32:18 +0000 From: Jarek Poplawski To: Patrick McHardy Cc: Karl Hiramoto , netdev@vger.kernel.org, netfilter@vger.kernel.org, LKML Subject: [PATCH] Re: problem with IPoA (CLIP), NAT, and VLANS Message-ID: <20090217093218.GA5341@ff.dom.local> References: <20090216232016.GA4803@ami.dom.local> <499A7D7D.1090607@trash.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <499A7D7D.1090607@trash.net> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Feb 17, 2009 at 10:03:57AM +0100, Patrick McHardy wrote: > Jarek Poplawski wrote: >> Very nice debugging, but I think your patch doesn't look like enough: >> 1) it skips copy for cloned skbs, >> 2) this skb_cow_head() is needed anyway, sometimes. >> So the real bug should be found in skb_cow_head() or elsewhere. >> >> I attach here 2 patches for testing: >> 1) skb->mac_header update: it looks like needed, but I don't know if >> it matters here, >> 2) an extention of your patch but with pskb_expand_head() called for >> one to one copy. >> >> BTW, if it's not a big problem it would be nice to try this e.g. on >> 2.6.29-rc5. > > The first patch looks fine. As for the second one, I would like > to understand why we're seing these packets. The VLAN driver uses > the original headroom plus the space it needs itself, which suggests > that the underlying driver specifies an incorrect headroom. > The driver doesn't appear to be part of the mainline kernel though. > The second patch is only for debugging: to check if it's about offset change or some values corrupted btw. Since the first patch looks quite natural I resend it for applying. I hope you'll ack it when Karl sends "Tested-by:" after checking it at least with the currently working (IPoE) vlan case. Thanks, Jarek P. -----------------> vlan: Update skb->mac_header in __vlan_put_tag(). After moving mac addresses in __vlan_put_tag() skb->mac_header() needs to be updated. Reported-by: Karl Hiramoto Signed-off-by: Jarek Poplawski --- include/linux/if_vlan.h | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/include/linux/if_vlan.h b/include/linux/if_vlan.h index f8ff918..e1ff5b1 100644 --- a/include/linux/if_vlan.h +++ b/include/linux/if_vlan.h @@ -210,6 +210,7 @@ static inline struct sk_buff *__vlan_put_tag(struct sk_buff *skb, u16 vlan_tci) /* Move the mac addresses to the beginning of the new header. */ memmove(skb->data, skb->data + VLAN_HLEN, 2 * VLAN_ETH_ALEN); + skb->mac_header -= VLAN_HLEN; /* first, the ethernet type */ veth->h_vlan_proto = htons(ETH_P_8021Q);