From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: Printing the driver name as part of the netdev watchdog message Date: Tue, 08 Jul 2008 14:57:38 -0700 (PDT) Message-ID: <20080708.145738.12692130.davem@davemloft.net> References: <4873BC67.8010205@opengridcomputing.com> <20080708.143158.182627862.davem@davemloft.net> <20080708144725.5b663d19@infradead.org> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: swise@opengridcomputing.com, rdreier@cisco.com, shemminger@vyatta.com, netdev@vger.kernel.org To: arjan@infradead.org Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:41202 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751120AbYGHV5i (ORCPT ); Tue, 8 Jul 2008 17:57:38 -0400 In-Reply-To: <20080708144725.5b663d19@infradead.org> Sender: netdev-owner@vger.kernel.org List-ID: From: Arjan van de Ven Date: Tue, 8 Jul 2008 14:47:25 -0700 > On Tue, 08 Jul 2008 14:31:58 -0700 (PDT) > David Miller wrote: > > > From: Steve Wise > > Date: Tue, 08 Jul 2008 14:13:43 -0500 > > > > > So should I put the rtnl acquisitions back in iw_cxgb3? Or fix > > > cxgb3 to not assume rtnl is held? > > > > It appears that the kernel always invokes the ethtool ops with > > RTNL held, and therefore that's the environment that needs > > to be ensured. > > so for the patch that started all this... do I need to go back to the > original where I didn't use ethool? > (since this is called from irq context....) What we need instead is to cache the info block into the netdev struct when the driver is ->open()'d, and then you can fetch it out of there however you like. Or, at least, that is one possible approach.