From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Samudrala, Sridhar" Subject: Re: [PATCH net-next 13/18] switchdev: remove unused NETIF_F_HW_SWITCH_OFFLOAD Date: Mon, 30 Mar 2015 13:48:34 -0700 Message-ID: <5519B6A2.4040204@intel.com> References: <1427704836-8776-1-git-send-email-sfeldma@gmail.com> <1427704836-8776-14-git-send-email-sfeldma@gmail.com> <551951CD.1060301@cumulusnetworks.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, jiri@resnulli.us, linux@roeck-us.net, f.fainelli@gmail.com To: roopa , sfeldma@gmail.com Return-path: Received: from mga03.intel.com ([134.134.136.65]:51910 "EHLO mga03.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751029AbbC3Usf (ORCPT ); Mon, 30 Mar 2015 16:48:35 -0400 In-Reply-To: <551951CD.1060301@cumulusnetworks.com> Sender: netdev-owner@vger.kernel.org List-ID: On 3/30/2015 6:38 AM, roopa wrote: > On 3/30/15, 1:40 AM, sfeldma@gmail.com wrote: >> From: Scott Feldman >> >> Flag is no longer needed so remove it. Using the attr get/set >> recurse algo >> obsoletes the flag. Setting the flag on bond/team interface, even >> when the >> consitient member ports didn't have flag set was confusing. >> >> Signed-off-by: Scott Feldman > > The flag was there to avoid the recursive lowerdev walk where possible. > bond will have the flag if any one slave can have it. You don't walk > the bond if it does not have the flag. > Also, this allows the user to disable the feature from userspace if > required and move everything to software. > I was also planning to use this flag to enable lowerdev walk for fdb add/delete on a team device with switchdev member ports and not attached to a bridge. Will submit a RFC patch for review. Thanks Sridhar