From mboxrd@z Thu Jan 1 00:00:00 1970 From: Or Gerlitz Subject: Re: zero features for a vlan over bond / vlan features Date: Thu, 16 Apr 2009 17:11:32 +0300 Message-ID: <49E73C94.5030009@Voltaire.com> References: <6223.1239113811@death.nxdomain.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org To: Jay Vosburgh , David Miller Return-path: Received: from fwil.voltaire.com ([193.47.165.2]:25953 "EHLO exil.voltaire.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752484AbZDPOLo (ORCPT ); Thu, 16 Apr 2009 10:11:44 -0400 In-Reply-To: <6223.1239113811@death.nxdomain.ibm.com> Sender: netdev-owner@vger.kernel.org List-ID: Jay Vosburgh wrote: > The features system now has a dev->vlan_features that lists the > features that will work through a vlan; bond_compute_features isn't > using netdev_increment/fix_features to additionally compute the vlan_features, > so that's ending up always empty even if the underlying device supports vlan passthrough. > I'm working on a patch; I'll see what I can come up with. Great, I will be happy to test the patch once you have it... > Note, however, that at this writing, the set of drivers that explicitly > supports VLAN passthrough (a non-zero vlan_features) is rather limited. Please correct me if I'm wrong - would be it correct to say that there's a concept of "VLAN acceleration" which comes into play when the NIC device advertises NETIF_F_HW_VLAN_xxx features and a concept of "VLAN passthrough" which means the NIC device have a non empty vlan_features bit mask? Looking on Linus tree I noted that indeed only few drivers have non-empty vlan_features and I wonder if the reason for that is indeed lack of HW support or just the driver maintainers being somehow slow to catch up with the vlan_features concept, Dave - any insight here? Or.