From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Herrenschmidt Subject: Re: Linux 2.6.18-rc6 Date: Wed, 06 Sep 2006 10:47:33 +1000 Message-ID: <1157503653.22705.156.camel@localhost.localdomain> References: <20060905122656.GA3650@aepfle.de> <1157490066.3463.73.camel@mulgrave.il.steeleye.com> <1157494837.22705.151.camel@localhost.localdomain> <1157498567.3463.91.camel@mulgrave.il.steeleye.com> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from gate.crashing.org ([63.228.1.57]:4802 "EHLO gate.crashing.org") by vger.kernel.org with ESMTP id S1751535AbWIFAsE (ORCPT ); Tue, 5 Sep 2006 20:48:04 -0400 In-Reply-To: <1157498567.3463.91.camel@mulgrave.il.steeleye.com> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: James Bottomley Cc: Olaf Hering , Linus Torvalds , linux-scsi@vger.kernel.org, Linux Kernel Mailing List On Tue, 2006-09-05 at 18:22 -0500, James Bottomley wrote: > On Wed, 2006-09-06 at 08:20 +1000, Benjamin Herrenschmidt wrote: > > Yes, it's a PCI error. > > Thanks, and the cat of /proc/scsi_host/host/signalling? > > My suspicion is the register doesn't actually exist on this card so it > doesn't actually respond on the bus. However, on my equivalent > everything works; largely I think because the only PC's I have don't > know how to signal a PCI error. Olaf will tell us (I don't have the hardware) but it's indeed a typical thing .... From my experience, Adaptec tend to very easily throw PCI aborts at you if it doesn't like something in a register access (which isn't necessarily a bad thing btw, other vendors are more lenient tho), and Macs are well known to Machine Check the box when getting a PCI error while most x86 boxes silently ignore them... Ben.