From owner-freebsd-scsi Wed Jan 5 12:48:30 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from front7.grolier.fr (front7.grolier.fr [194.158.96.57]) by hub.freebsd.org (Postfix) with ESMTP id C401C155C2 for ; Wed, 5 Jan 2000 12:48:25 -0800 (PST) (envelope-from groudier@club-internet.fr) Received: from localhost (ppp-163-168.villette.club-internet.fr [195.36.163.168]) by front7.grolier.fr (8.9.3/No_Relay+No_Spam_MGC990224) with SMTP id VAA02090; Wed, 5 Jan 2000 21:48:04 +0100 (MET) Date: Wed, 5 Jan 2000 22:15:22 +0100 (MET) From: Gerard Roudier X-Sender: groudier@localhost To: Adam Cc: scsi@FreeBSD.ORG Subject: Re: sym troubles In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Tue, 4 Jan 2000, Adam wrote: > On Wed, 5 Jan 2000, Gerard Roudier wrote: >=20 > >I suggest you the following magic command that should avoid the offendin= g > >CHECK CONDITION and should allow the driver to negotiate SYNC=3D20MHz on= =20 > >the normal INQUIRY command and prove us it did so as expected: > > > > ******************************** > > * camcontrol inquiry cd1 -D -R * > > ******************************** > > > >I expect this one to finally tell us the truth. Answer should resemble t= he > >following as you know: > > > >pass1: Removable CD-ROM SCSI-2 device=20 > >pass1: 40.000MB/s transfers (16bit) + some BLAH BLAH BLAH > > > > G=E9rard. >=20 > True! :) Does this mean it is operating properly a) without > special prodding b) only with prodding c) never? It means that I haven't been wrong a single time in my explanation of the= =20 apparent trouble. > I'm just curious if any of this would bring down the bus speed when > accessing the cdrom if I had any UW hd's on the bus for example, or if it > is purely a cosmetic issue. Also wondering why ncr0 works fine with it := P The sym driver will give you full speed for the transfer of useful data for all the devices on the BUS. Justin suggested me the right thing to do by private email. Anyway, both the ncr and the sym will operate at 40 MB/second when transferring data.=20 The CAM/SCSI/camcontrol just announce the speed after a CHECK CONDITION and this is the only time where data transfer is turned to ASYNC by sym.=20 The sym _does_ negotiate SYNC on the next command, but you are not notifying of that. By the way both the ncr and the sym require 2 SCSI commands for negotiating WIDE+SYNC (kind of legacy :) ), but the ncr does not renegotiate WIDE on AUTO-SENSE and this makes the difference you observe.= =20 The aic7xxx does all the work (WIDE+SYNC) in a single SCSI command as=20 Justin Gibbs suggested me to implement in the sym driver. =20 The change is on my todo list for the sym. But I will not provide an immediate change given that this will not improve anything in essence and= =20 does not affect either stability or speed. Now, if you donnot beleive me, you have the choice to use the ncr instead. You CD/ROM will be equally fast, in my opinion. Thanks for your problem report, G=E9rard. PS: I donnot see what may still be unclear about this topic. So, if something is still unclear for you, private email should be more reasonnable. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message