From nick@cecomputing.co.uk Sun Sep 27 13:52:37 2009 Received: from moutng.kundenserver.de ([212.227.126.187]) by stoneboat.aleph1.co.uk with esmtp (Exim 4.69) (envelope-from ) id 1MrtEu-0003SU-Eq for balloon@balloonboard.org; Sun, 27 Sep 2009 13:52:37 +0100 Received: from localhost (host81-130-93-71.in-addr.btopenworld.com [81.130.93.71]) by mrelayeu.kundenserver.de (node=mreu1) with ESMTP (Nemesis) id 0LzWb8-1MMb3Q2yry-014j2k; Sun, 27 Sep 2009 14:52:26 +0200 Received: from a64.local ([192.168.40.40]) by localhost with esmtp (Exim 4.69) (envelope-from ) id 1MrtEn-0002Vt-Qx for balloon@balloonboard.org; Sun, 27 Sep 2009 13:52:25 +0100 Message-ID: <4ABF5FEC.1070906@cecomputing.co.uk> Date: Sun, 27 Sep 2009 13:51:56 +0100 From: Nick Bane User-Agent: Thunderbird 2.0.0.23 (Windows/20090812) MIME-Version: 1.0 To: "balloon@balloonboard.org" Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Provags-ID: V01U2FsdGVkX18BDlDuqInwbuZvQ+hWwPUwyFWN/pqc+OezXDn +SCox2b59hWFE6h1herphpawm7N6sAT/nFfJIVYlxaN0FrgBiD OoDSm8+oo0IdZ2ZfPTxwQ== X-SA-Exim-Connect-IP: 212.227.126.187 X-SA-Exim-Mail-From: nick@cecomputing.co.uk X-Spam-Checker-Version: SpamAssassin 3.2.5 (2008-06-10) on stoneboat.aleph1.co.uk X-Spam-Level: X-Spam-Status: No, score=-2.5 required=4.5 tests=AWL,BAYES_20,SPF_HELO_PASS autolearn=no version=3.2.5 X-SA-Exim-Version: 4.2.1 (built Wed, 25 Jun 2008 17:14:11 +0000) X-SA-Exim-Scanned: Yes (on stoneboat.aleph1.co.uk) Subject: [Balloon] udc updates X-BeenThere: balloon@balloonboard.org X-Mailman-Version: 2.1.11 Precedence: list List-Id: Balloon List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 27 Sep 2009 12:52:37 -0000 I just updated the udc support for g_ether gadget. This is in 2 parts and I thought it needed some explanation. Firstly ether gadget support for suspend is added and implemented for the cdc_subset driver. Suspend in pxa270-speak is where the gadget loses its host connection for 3ms or more. As the balloon3 has no hardware detection for cable plug/unplug, this is one reliable way of detecting the cable being unplugged and thus notifying the ethernet link layer that no more traffic is possible. The second part is a horrible hack on the back of the first part. I failed to get ifpugd to monitor the state of the interface even when ethtool link detection was routed to netif_carrier. Maybe that could be persuaded to work and would be a better solution that what follows. I added a usermode helper that explicitly (yes it should be via an installable proxy but time etc) calls ifup/down when the carrier is said to come and go. Note that this will happen in interrupt context so we must defer processing till later. Also note that there is whole slew if carrier present/absent calls during connect/disconnect so I run a cumulative timeout to only request one ifup/down per flurry of activity and then schedule the helper function. And it seems to work ok too. Nick Bane