From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: BusLogic driver reports non-existent disks since VMware 1.0.4 Date: Sun, 16 Mar 2008 16:56:18 -0500 Message-ID: <1205704578.6767.113.camel@localhost.localdomain> References: <47DD834B.9050807@gentoo.org> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from accolon.hansenpartnership.com ([76.243.235.52]:51480 "EHLO accolon.hansenpartnership.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752436AbYCPV4W (ORCPT ); Sun, 16 Mar 2008 17:56:22 -0400 In-Reply-To: <47DD834B.9050807@gentoo.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Daniel Drake Cc: linux-scsi@vger.kernel.org, lnz@dandelion.com On Sun, 2008-03-16 at 20:30 +0000, Daniel Drake wrote: > There's no evidence that this is a BusLogic driver bug (as opposed to > being a VMware bug), but as I couldn't find any mention on this list I > thought I'd bring it to peoples attention: > > If you upgrade from vmware 1.0.3 to 1.0.4, when using the BusLogic > emulation (as opposed to LSI logic), the kernel reports existence of > many disks that do not actually exist. Accessing those disks obviously > causes hangs and error messages. > > http://communities.vmware.com/thread/103939?tstart=0&start=0 > https://bugs.gentoo.org/show_bug.cgi?id=198296 > > dmesg output from such a system: > https://bugs.gentoo.org/attachment.cgi?id=135371&action=view > > This is similar but not the same as another vmware vs scsi driver problem: > http://marc.info/?l=linux-scsi&m=118590789115292&w=2 > (that one is different and has been fixed in vmware) > > If anyone is interested in investigating, I can pass on patches to > people who can test them. It does rather look like a vmware emulation problem. The Short INQUIRY read: scsi 0:0:0:0: Direct-Access VMware, VMware Virtual S 1.0 PQ: 0 ANSI: 2 scsi scan: INQUIRY result too short (5), using 36 scsi 0:0:1:0: Direct-Access PQ: 0 ANSI: 0 Implies the inquiry returned all zeros (length of 5 is because that's the offset into the INQUIRY of the additional length), which is also a strong indicator of a vmware issue. There have been reports of BusLogic Flashpoint working, but that's practically a separate driver. The only real way to be sure is for someone to try it out with a real BT-958 and see. James