From mboxrd@z Thu Jan 1 00:00:00 1970 From: Willem Riede Subject: Re: [PATCH] switch scsi upper driver probing to the driver model Date: Fri, 16 May 2003 22:25:07 -0400 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20030517022507.GA1480@linnie.riede.org> References: <20030516182039.A7369@lst.de> <20030516235718.GA3260@beaverton.ibm.com> Reply-To: wrlk@riede.org Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII; Content-Transfer-Encoding: 7BIT Return-path: Received: from pcp02781107pcs.eatntn01.nj.comcast.net ([68.85.61.149]:5359 "EHLO linnie.riede.org") by vger.kernel.org with ESMTP id S264649AbTEQCKD (ORCPT ); Fri, 16 May 2003 22:10:03 -0400 Received: from linnie.riede.org (localhost.localdomain [127.0.0.1]) by linnie.riede.org (8.12.8/8.12.8) with ESMTP id h4H2P7Y8002289 for ; Fri, 16 May 2003 22:25:07 -0400 Content-Disposition: inline In-Reply-To: <20030516235718.GA3260@beaverton.ibm.com>; from andmike@us.ibm.com on Fri, May 16, 2003 at 19:57:18 -0400 List-Id: linux-scsi@vger.kernel.org To: linux-scsi@vger.kernel.org On 2003.05.16 19:57, Mike Anderson wrote: > > I ran your patch and I dead locked on st_probe. This is due to the code > in st_probe and osst_probe that do device registers on the > sdev_driverfs_dev.bus which is the bus that the probe is being called > from already. I had an old patch that removed these before due the driver > model requirement that objects on the device list must be of one type. > At the time I did not have a place to move these attributes. > > We now have /bus/scsi/drivers/{uld} for driver attributes. We also could > create a class for other attributes if needed. We can also use udev for > the creation of special nodes. > > ftp://ftp.kernel.org/pub/linux/utils/kernel/hotplug/ > > In the short term is the st and osst maintainers are ok with it we could > remove them and then create them in there proper home. As the osst maintainer I concur. I am going to need a class, I believe, as I want to export some info from the tape header, i.e. per device (very useful for me for support purposes). So some re-arrangement is going to happen anyway. Thanks, Willem Riede.