From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: AHCI driver preferring nr_ports over port map Date: Mon, 04 Feb 2008 22:16:44 +0900 Message-ID: <47A7103C.1040804@gmail.com> References: <47A70EA90200007800045D32@public.id2-vpn.continvity.gns.novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Return-path: Received: from an-out-0708.google.com ([209.85.132.242]:19217 "EHLO an-out-0708.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751881AbYBDNQu (ORCPT ); Mon, 4 Feb 2008 08:16:50 -0500 Received: by an-out-0708.google.com with SMTP id d31so520102and.103 for ; Mon, 04 Feb 2008 05:16:49 -0800 (PST) In-Reply-To: <47A70EA90200007800045D32@public.id2-vpn.continvity.gns.novell.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Jan Beulich Cc: jgarzik@pobox.com, linux-ide@vger.kernel.org Jan Beulich wrote: >> Do you have any real case where the above behavior causes problem? > > It's not strictly a problem (i.e. nothing really mis-behaves), but it made > me wonder why the box I saw this on gets 6 ahci device instances set > up when spec as well as port map say there ought to be only 5. After > looking at the ESB2 spec it seemed that behavior was clearly violating > the spec: "For ports that are not available, software must not read or > write to registers within that port.", which contradicts status being > displayed for (and therefore status being read) from the 6th (not > present) port. Well, two values don't agree with each other and we know for a fact that vendors sometimes get PI wrong, so we trust n_ports in such cases. We can reverse the behavior but that's likely to cause more problems than it fixes. -- tejun