From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [PATCH]: Flexible timeout infrastructure Date: Tue, 15 Jun 2004 16:49:13 +0100 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20040615154913.GA13917@infradead.org> References: <40CF0F9F.4050902@adaptec.com> <1087313241.2710.40.camel@laptop.fenrus.com> <40CF1852.6030306@adaptec.com> <20040615154223.GA13780@infradead.org> <40CF19C0.5080507@adaptec.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from [213.146.154.40] ([213.146.154.40]:2688 "EHLO pentafluge.infradead.org") by vger.kernel.org with ESMTP id S265716AbUFOPtP (ORCPT ); Tue, 15 Jun 2004 11:49:15 -0400 Content-Disposition: inline In-Reply-To: <40CF19C0.5080507@adaptec.com> List-Id: linux-scsi@vger.kernel.org To: Luben Tuikov Cc: Christoph Hellwig , arjanv@redhat.com, SCSI Mailing List On Tue, Jun 15, 2004 at 11:46:08AM -0400, Luben Tuikov wrote: > >That's why the plan is to move it into the transport class slowly. Remeber > >there are much more SPI drivers than just aic7xxx/aic79xx ;-) > > Yes, in which case it would look like FreeBSD. Then SCSI Core will > have to implement _all_ transports it wants to support (a la USB Storage). Umm, no. Look at a recent Linux 2.6 codebase. We support modular transport classes so it's _NOT_ a big mess like FreeBSD. And we won't stop providing sensible defaults abyway. > But if we *do* allow for this patch, then _moving_ this functionality > into the transport class recovery would be really easy. Just move > eh_cmd_timed_out() into the transport class. Show us a patch with your SPI EH improvements.