From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Cooper Subject: Re: [PATCH] AMD/guest_iommu: Disable guest iommu support Date: Mon, 6 Oct 2014 10:11:06 +0100 Message-ID: <54325CAA.80909@citrix.com> References: <1412258549-18671-1-git-send-email-andrew.cooper3@citrix.com> <54323AC3.5050200@amd.com> <5432681F020000780003C541@mail.emea.novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <5432681F020000780003C541@mail.emea.novell.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Suravee Suthikulpanit Cc: Roberto Luongo , Aravind Gopalakrishnan , Jan Beulich , Xen-devel List-Id: xen-devel@lists.xenproject.org On 06/10/14 08:59, Jan Beulich wrote: >>>> On 06.10.14 at 08:46, wrote: >> The function guest_iommu_set_base() was added by Wei before my time. I >> checked the log and didn't see the evidence that this code has ever been >> used. Do you remember why he added this code to begin with, and how this >> was planned to be used? > The plan here clearly was to allow guests to see a virtual IOMMU, > and the setting of the respective base address would then be > a requirement for tool stack and/or virtual firmware to set up the > environment for the guest. > > Jan > The commit which introduces the virtual IOMMU (c/s 4547137c3) indicates that a virtual IOMMU is necessary for correct functioning of ATS-capable guest device drivers using PASID or PRI, in passthrough. ~Andrew