From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Woodhouse Subject: Re: [PATCH] phylib: Add autoload support for the LXT973 phy. Date: Wed, 23 Jun 2010 10:00:12 +0100 Message-ID: <1277283612.16262.195.camel@macbook.infradead.org> References: <20100531130932.GA15845@riccoc20.at.omicron.at> <1277210293.21798.11.camel@localhost> <20100623053723.GA3404@riccoc20.at.omicron.at> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org To: Richard Cochran Return-path: Received: from casper.infradead.org ([85.118.1.10]:57956 "EHLO casper.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752256Ab0FWJAO (ORCPT ); Wed, 23 Jun 2010 05:00:14 -0400 In-Reply-To: <20100623053723.GA3404@riccoc20.at.omicron.at> Sender: netdev-owner@vger.kernel.org List-ID: On Wed, 2010-06-23 at 07:37 +0200, Richard Cochran wrote: > > Question about the whole PHY MODULE_DEVICE_TABLE system: > > I recently posted a phy driver for the National Semiconductor > DP83640. During development, I used drivers/net/arm/ixp4xx_eth.c as > the MAC driver, which was linked into the kernel (not a module). I > noticed that the phy driver's probe function only gets called if the > phy driver is also statically linked, but not when it is loaded as a > module. > > Is this the correct behavior? Hm, that seems like the _expected_ behaviour, certainly. The MAC driver will probe its device at boot time, and will issue a request_module() to load the a specific PHY driver if there is one. When no such module turns up (which it won't if you have no file system mounted yet), it'll just fall back to the generic PHY support. -- David Woodhouse Open Source Technology Centre David.Woodhouse@intel.com Intel Corporation