From mboxrd@z Thu Jan 1 00:00:00 1970 From: Anton Blanchard Subject: Re: [PATCH] serialize bus scanning Date: Thu, 28 Aug 2003 08:59:40 +1000 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20030827225940.GC7852@krispykreme> References: <20030825122416.GE15506@lst.de> <20030827202908.GB7852@krispykreme> <20030827223250.GB2664@beaverton.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from dp.samba.org ([66.70.73.150]:62669 "EHLO lists.samba.org") by vger.kernel.org with ESMTP id S262476AbTH0W7s (ORCPT ); Wed, 27 Aug 2003 18:59:48 -0400 Content-Disposition: inline In-Reply-To: <20030827223250.GB2664@beaverton.ibm.com> List-Id: linux-scsi@vger.kernel.org To: Christoph Hellwig , James Bottomley , linux-scsi@vger.kernel.org > Since I have not seen your tree I am speaking with lack of knowledge, but > it seems odd that your device tree would contain spinup knowledge per > host. The power issue you are trying to avoid is per disk enclosure > / cabinet. A per host value seems like it could still lead to a power > issue given a large enough configuration. Yeah you are right, its currently a global value. The firmware knows power domains so it could get finer grained but it would be at the host bridge level. Anyway all this intelligence to parse this stuff should be in an initramfs program talking to sysfs. I was quite pleased with the sysfs rescan functionality recently, we had someone kick the power on a scsi drawer and were able to get linux to rescan and spin up the disks again without rebooting. Anton