From mboxrd@z Thu Jan 1 00:00:00 1970 From: Richard Cochran Subject: Re: [PATCH] net-timestamp: Fix a documentation typo Date: Tue, 25 Nov 2014 09:33:42 +0100 Message-ID: <20141125083342.GA4411@localhost.localdomain> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Andy Lutomirski , Network Development , "David S. Miller" To: Willem de Bruijn Return-path: Received: from mail-wi0-f182.google.com ([209.85.212.182]:34719 "EHLO mail-wi0-f182.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750963AbaKYIds (ORCPT ); Tue, 25 Nov 2014 03:33:48 -0500 Received: by mail-wi0-f182.google.com with SMTP id h11so565672wiw.9 for ; Tue, 25 Nov 2014 00:33:46 -0800 (PST) Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Mon, Nov 24, 2014 at 05:38:21PM -0500, Willem de Bruijn wrote: > The datagram feature existed before I added the counter and stream > support, so returning the entire packet in that case is legacy > behavior, I suppose. I did not intend to expose network headers for > the new stream socket interface, though. I suppose that idea (not mine) behind returning the entire frame was to allow the user to send multiple frames, and then figure out which time stamp goes with which sent frame. The whole thing is weak for two reasons, firstly you aren't able to send two identical frames, and secondly there are quite a few time stamping cards out there which only handle one transmit time stamp at a time. Thus the ptp4l application always waits for the hardware time stamp before sending another frame and never looks at the returned frame data. Thanks, Richard