From mboxrd@z Thu Jan 1 00:00:00 1970 From: Doug Ledford Subject: Re: [PATCH + RFC] Beginning of some updates to scsi mid layer Date: Mon, 1 Jul 2002 15:59:28 -0400 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20020701155928.H776@redhat.com> References: <20020701151553.G776@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: ; from mjacob@feral.com on Mon, Jul 01, 2002 at 12:23:56PM -0700 List-Id: linux-scsi@vger.kernel.org To: Matthew Jacob Cc: =?iso-8859-1?Q?G=E9rard_Roudier?= , Jeremy Higdon , Martin Peschke3 , Pete Zaitcev , linux-scsi@vger.kernel.org On Mon, Jul 01, 2002 at 12:23:56PM -0700, Matthew Jacob wrote: > Fibre. Devices 'go away' for shortish to longish periods of time on Fibre (or > iSCSI for that matter). Lacking a sane framework that is USB-like about > hotpluggable that does *not* cause complete system spasm when this happens > (and it happens a lot), what one wants is a bit of hysteresis to give the > disks that went away a chance to log back into the fabric and thus appear > again (as if they had never gone). (I knew this, but it had slipped my mind). This reminds me of something I was thinking about for the mid layer code stuff. Does FreeBSD have it's drivers declare their connection method anywhere? It seems to me there are a limited number of connection specific details that the mid layer would benefit from being aware of (for example > 8 luns on SCSI-2 fiber devices, but not on SCSI-2 SPI devices, the selection timeout issue you just raised with transient disk loss which should not happen on SPI, things like that). Does the FreeBSD mid layer do anything with that sort of information? -- Doug Ledford 919-754-3700 x44233 Red Hat, Inc. 1801 Varsity Dr. Raleigh, NC 27606