From mboxrd@z Thu Jan 1 00:00:00 1970 From: Auger Eric Subject: Re: [PATCH v7 09/17] KVM: arm64: introduce ITS emulation file with MMIO framework Date: Mon, 4 Jul 2016 16:15:38 +0200 Message-ID: <9fea837a-d45d-ae2b-6442-0a89db124946@redhat.com> References: <20160628123230.26255-1-andre.przywara@arm.com> <20160628123230.26255-10-andre.przywara@arm.com> <3fec9632-ce39-3508-2b71-4049fc959a64@redhat.com> <32d54016-731d-e4a9-272d-8b0934ebec31@arm.com> <8546ac45-f7cd-31af-a261-bb85ca7305ca@redhat.com> <33ff8f2c-89e0-5778-71a6-9d116e061752@arm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Cc: kvmarm@lists.cs.columbia.edu, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org To: Andre Przywara , Marc Zyngier , Christoffer Dall Return-path: Received: from mx1.redhat.com ([209.132.183.28]:59912 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753025AbcGDOPm (ORCPT ); Mon, 4 Jul 2016 10:15:42 -0400 In-Reply-To: <33ff8f2c-89e0-5778-71a6-9d116e061752@arm.com> Sender: kvm-owner@vger.kernel.org List-ID: On 04/07/2016 16:00, Andre Przywara wrote: > Hi, > > On 04/07/16 14:54, Auger Eric wrote: >> Hi Andre, >> >> On 04/07/2016 15:38, Andre Przywara wrote: >>> 2) on KVM_DEV_ARM_VGIC_GRP_ADDR check that "initialized" is false >>> 3) on KVM_DEV_ARM_VGIC_CTRL_INIT check that "initialized" is false >>> 4) after KVM_DEV_ARM_VGIC_CTRL_INIT set initialized to true >>> >>> I don't want to rely solely on vgic_its_base being ADDR_UNDEF, since I >>> expect we extend that interface later by supporting more options (for >>> instance to size the virtual ITS, in terms of devices, number of LPIs, >>> collections, etc.) >>> >>> Does that make sense? >>> Or do we need more here? >> >> So you will accept a call to KVM_DEV_ARM_VGIC_CTRL_INIT with an unset >> address? I don't think this is relevant. > > Ah, good point. Will fix this. > > However I think we do the same on GICv2/GICv3 at the moment. CTRL_INIT > will succeed, I guess we fail the first KVM_RUN then, right? Yes the map_resources will fail with UNDEF addresses Eric > > Cheers, >