From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jiri Benc Subject: Re: Delivery Status Notification (Failure) Date: Thu, 24 Sep 2015 15:09:06 +0200 Message-ID: <20150924150906.067bba62@griffin> References: <047d7b5dbb76a0065405207cfe24@google.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org To: Naveen Return-path: Received: from mx1.redhat.com ([209.132.183.28]:33487 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754954AbbIXNJJ (ORCPT ); Thu, 24 Sep 2015 09:09:09 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Thu, 24 Sep 2015 08:03:39 -0400, Naveen wrote: > Right now the only way to detect if an interface has a peer(veth) > is touse ethtool -S. Instead I was thinking of adding it to device > attributes as a peer_ifndex entry. In drivers/net/veth.c I could add > an entry to > dev->groups for a veth specific group. Would that be acceptable ? Recent kernels announce the peer in IFLA_LINK, including its netnsid (thus it's finally possible to track peers across net namespaces). Just try "ip a" with a recent kernel. Jiri -- Jiri Benc