From mboxrd@z Thu Jan 1 00:00:00 1970 From: Joerg Roedel Subject: Re: Kernel Oops: iommu related? Date: Fri, 13 Feb 2015 23:27:01 +0100 Message-ID: <20150213222701.GH29106@8bytes.org> References: <54DCE8A6.4000608@compro.net> <20150212180846.GD29106@8bytes.org> <54DCF024.30309@compro.net> <54DD00FE.2050009@compro.net> <20150212211340.GF29106@8bytes.org> <54DE07DA.9050703@compro.net> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <54DE07DA.9050703-n2QNKt385d+sTnJN9+BGXg@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: iommu-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org Errors-To: iommu-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org To: Mark Hounschell Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org List-Id: iommu@lists.linux-foundation.org On Fri, Feb 13, 2015 at 09:19:06AM -0500, Mark Hounschell wrote: > OK. Attached. Thanks, the IVRS table looks good. Device 0f.00.0 is listed there as the alias target for a group of other devices: > [ 1.405719] AMD-Vi: DEV_ALIAS_RANGE devid: 0f:01.0 flags: 00 devid_to: 0f:00.0 > [ 1.405722] AMD-Vi: DEV_RANGE_END devid: 12:1f.7 > BTW when I blacklist the aic7xxx driver [last > unloaded: aic7xxx] I can load and unload any module all day long. Okay, this driver seems to be the problem, it has a bug somewhere in using the DMA-API. This causes the iommu problems. The device it serves is behind the alias range for 0f:00.0, so any faults from that device will show up as coming from 0f:00.0, like in the io-page-faults you see in dmesg. Joerg