From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753381Ab2AZSZo (ORCPT ); Thu, 26 Jan 2012 13:25:44 -0500 Received: from db3ehsobe001.messaging.microsoft.com ([213.199.154.139]:16090 "EHLO DB3EHSOBE001.bigfish.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751660Ab2AZSZn (ORCPT ); Thu, 26 Jan 2012 13:25:43 -0500 X-SpamScore: -11 X-BigFish: VS-11(zzbb2dI9371I1432N98dKzz1202hzzz2dhc1bhc31hc1ah2a8h668h839h) X-Forefront-Antispam-Report: CIP:70.37.183.190;KIP:(null);UIP:(null);IPV:NLI;H:mail.freescale.net;RD:none;EFVD:NLI Message-ID: <4F219A9C.8000400@freescale.com> Date: Thu, 26 Jan 2012 12:25:32 -0600 From: Scott Wood User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:6.0.2) Gecko/20110906 Thunderbird/6.0.2 MIME-Version: 1.0 To: Joerg Roedel CC: Sethi Varun-B16395 , "iommu@lists.linux-foundation.org" , Ohad Ben-Cohen , Tony Lindgren , Hiroshi DOYU , "linux-kernel@vger.kernel.org" , Laurent Pinchart , Wood Scott-B07421 , David Brown , David Woodhouse Subject: Re: [PATCH 2/5] iommu/amd: Implement DOMAIN_ATTR_GEOMETRY attribute References: <1326983405-319-1-git-send-email-joerg.roedel@amd.com> <1326983405-319-3-git-send-email-joerg.roedel@amd.com> <20120120160344.GG2205@amd.com> In-Reply-To: <20120120160344.GG2205@amd.com> Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit X-OriginatorOrg: freescale.com Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 01/20/2012 10:03 AM, Joerg Roedel wrote: > On Thu, Jan 19, 2012 at 05:16:48PM +0000, Sethi Varun-B16395 wrote: >>> struct iommu_domain { >>> struct iommu_ops *ops; >>> void *priv; >>> iommu_fault_handler_t handler; >>> + struct iommu_domain_geometry geometry; >>> }; >> In case of our iommu implementation the iommu_domain_geometry would >> include additional attributes. Why can't we let the geometry be >> Implementation dependent? > > Because the values in this generic geometry-struct make sense for more > than one IOMMU (short-term for Tegra GART and Freescale PAMU). If you > have additional vendor-specific ones you could add them via a > vendor-specific attribute. Freescale PAMU is the IOMMU that Varun is talking about, that needs additional geometry attributes (in particular, subwindow count). How should a PAMU driver interpret "force_aperture"? When would DMA ever be allowed outside the specified range? What does the range mean in that case? -Scott