From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: Bugs in multipath scsi in 4.3-rc2 Date: Fri, 25 Sep 2015 10:31:18 -0700 Message-ID: <1443202278.2188.13.camel@HansenPartnership.com> References: <20150925121636.GC12540@fergus.ozlabs.ibm.com> <20150925151802.GB20282@lst.de> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: Received: from bedivere.hansenpartnership.com ([66.63.167.143]:58129 "EHLO bedivere.hansenpartnership.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932474AbbIYRbj (ORCPT ); Fri, 25 Sep 2015 13:31:39 -0400 In-Reply-To: <20150925151802.GB20282@lst.de> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Christoph Hellwig Cc: Paul Mackerras , linux-scsi@vger.kernel.org On Fri, 2015-09-25 at 17:18 +0200, Christoph Hellwig wrote: > Hi Paul, > > can you send the request_module fix as a proper signed off and described > patch? I'll figure out what w can do about async scan vs request_module > in the meantime. So the warning seems to be because scsi_dh_find_driver() is not quite consistent. For everything except alua, it scans the dh driver list to see what might attach to the device. It returns "alua" if the TPGS field is anything other than zero, regardless of whether the alua driver is loaded. We could fix the problem by returning NULL if the alua driver isn't present ... would that have any other adverse consequences? James