From owner-freebsd-scsi Wed Dec 25 02:21:32 1996 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.4/8.8.4) id CAA04610 for freebsd-scsi-outgoing; Wed, 25 Dec 1996 02:21:32 -0800 (PST) Received: from irz301.inf.tu-dresden.de (irz301.inf.tu-dresden.de [141.76.1.11]) by freefall.freebsd.org (8.8.4/8.8.4) with SMTP id CAA04595 for ; Wed, 25 Dec 1996 02:21:28 -0800 (PST) Received: from sax.sax.de (sax.sax.de [193.175.26.33]) by irz301.inf.tu-dresden.de (8.6.12/8.6.12-s1) with ESMTP id LAA28791; Wed, 25 Dec 1996 11:21:18 +0100 Received: (from uucp@localhost) by sax.sax.de (8.6.12/8.6.12-s1) with UUCP id LAA18286; Wed, 25 Dec 1996 11:21:18 +0100 Received: (from j@localhost) by uriah.heep.sax.de (8.8.4/8.6.9) id LAA16758; Wed, 25 Dec 1996 11:05:03 +0100 (MET) From: J Wunsch Message-Id: <199612251005.LAA16758@uriah.heep.sax.de> Subject: Re: cvs commit: src/sys/scsi cd.c To: freebsd-scsi@FreeBSD.org (FreeBSD SCSI list) Date: Wed, 25 Dec 1996 11:05:03 +0100 (MET) Cc: msmith@atrad.adelaide.edu.au (Michael Smith) Reply-To: joerg_wunsch@uriah.heep.sax.de (Joerg Wunsch) In-Reply-To: <199612250117.LAA21192@genesis.atrad.adelaide.edu.au> from Michael Smith at "Dec 25, 96 11:47:48 am" X-Phone: +49-351-2012 669 X-PGP-Fingerprint: DC 47 E6 E4 FF A6 E9 8F 93 21 E0 7D F9 12 D6 4E X-Mailer: ELM [version 2.4ME+ PL17 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk (Moved to -scsi) As Michael Smith wrote: > > Modified: sys/scsi cd.c > > Slightly re-order the sequence of commands issued, so try to send > > the START UNIT command before testing whether the device is ready. > > Maybe it should be done even earlier, i'm not 100 % sure. > > >From recent experience with a box full of disks jumpered to spin up on > command (inadequate power supply), I'd really appreciate it if the > START UNIT command could be sent during the probe, followed by a > _long_ wait for the drive to become ready. But that's a little unrelated to this commit, right? This one was for CDs. However, it looks as if the original drivers for sd/cd/od were all cloned from the same, and thus inherited all the bugs... so i think the same problem still exists in the sd driver. Why do you want to have the disks spun up at probe time? IMHO, issuing the START UNIT at open time would be sufficient. There's no need to start the spindles before you're going to access them. The INQUIRY command is usually answered fine without spinning. Of course, the long wait is still required. The current timeout is 10 seconds, it's in the function scsi_start_unit() in sys/scsi/scsi_base.c. Feel free to bump this to 60 seconds once you know what you need. (I think you could experiment with a change to the sd driver that is similar to the cd driver change we're discussing about here.) -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)