From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bob Wirka Subject: Re: NFS and Network Driver Question Date: Tue, 19 Oct 2004 13:03:16 -0500 Sender: netdev-bounce@oss.sgi.com Message-ID: <417556E4.4090406@rtcworks.com> References: <41704198.8000206@rtcworks.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Cc: netdev@oss.sgi.com, linux-net@vger.kernel.org Return-path: In-Reply-To: <41704198.8000206@rtcworks.com> Errors-to: netdev-bounce@oss.sgi.com List-Id: netdev.vger.kernel.org Please disregard my last post, I believe the problem is solved. Thanks again to all, especially Thomas Graf. The essential things to learn from this exercise are: 1. A network driver compiled into the kernel must have an initialization function that allocates the 'net_device' structure and passes this to the kernel through the 'register_netdev()' call. You can reuse the module-version initialization function; rename it and prefix it's declaration with 'static __init'. 2. Add the 'module_init( init_fcn )' macro after the initialization function. This will cause the kernel to execute that code when booting. Onward and upward. Bob Wirka Realtime Control Works Bob Wirka wrote: > Hello, > > I'm trying to build a kernel that mounts a NFS root file system. This > is an embedded system; it uses an SMSC LAN91C111 network chip that is > hardwired to I/O addres 0x300 and IRQ 5. I've been using the driver > (as supplied by SMSC) as a module, and it works fine. Now I'm trying > to incorporate it into the kernel build so that I can mount an NFS > file system when the system boots. > > I've added the source code to the kernel tree, and modified the > Makefile(s) and Config.in files so that the driver <> to be > compiled into the kernel. The kernel will load, but the chip is not > initialized and the NFS fails. > > Since I don't know how to pass the io address and irq number to the > kernel, I've hard-coded them into the driver's init function (which I > don't see being called). > > If someone could point me in the right direction, I'd appreciate it. > > Thank you, > > Bob Wirka > Realtime Control Works > >