From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [PATCH 3/3] scsi: Implement 'ZBC' device type Date: Mon, 2 Jun 2014 02:20:53 -0700 Message-ID: <20140602092053.GD32301@infradead.org> References: <1401692478-23557-1-git-send-email-hare@suse.de> <1401692478-23557-4-git-send-email-hare@suse.de> <20140602083919.GA31170@infradead.org> <538C39EB.90109@suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from bombadil.infradead.org ([198.137.202.9]:40311 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752843AbaFBJUy (ORCPT ); Mon, 2 Jun 2014 05:20:54 -0400 Content-Disposition: inline In-Reply-To: <538C39EB.90109@suse.de> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Hannes Reinecke Cc: Christoph Hellwig , James Bottomley , linux-scsi@vger.kernel.org On Mon, Jun 02, 2014 at 10:46:35AM +0200, Hannes Reinecke wrote: > >But I'm uneasy with adding anything like this for now, for one the specs > >aren't even anywhere close to done, and second attaching ZBC devices > >in general doesn't sound like a very smart idea, we really need a > >strategy for the sequential required zones first. > > > In theory, yes. > > In practice, however, ZAC drives will show up a normal ATA devices > without that patch. Which is leading to even more trouble. > So with this patch we're at least signalling to the user that > something odd might happen. So how about we drop the sd.c changes so that no ULD gets attached, and keep the rest with the review comments addressed? That way we don't accidentally attach a ZBC/ZAC disk, but do get proper diagnostics and can even use SG_IO with the experimental ZBC or ZAC commands.