From mboxrd@z Thu Jan 1 00:00:00 1970 From: Cam Macdonell Subject: Re: [PATCH v2] Driver for Inter-VM shared memory device for KVM supporting interrupts. Date: Tue, 19 May 2009 10:51:08 -0600 Message-ID: <4A12E37C.700@cs.ualberta.ca> References: <1241713567-17256-1-git-send-email-cam@cs.ualberta.ca> <200905181607.32924.borntraeger@de.ibm.com> <4A117007.7040202@redhat.com> <200905191100.14252.borntraeger@de.ibm.com> <4A127776.9030400@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: =?UTF-8?B?Q2hyaXN0aWFuIEJvcm50csOkZ2Vy?= , kvm@vger.kernel.org, Rusty Russell , Christian Ehrhardt To: Avi Kivity Return-path: Received: from fleet.cs.ualberta.ca ([129.128.22.22]:58445 "EHLO fleet.cs.ualberta.ca" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751462AbZESQvI (ORCPT ); Tue, 19 May 2009 12:51:08 -0400 In-Reply-To: <4A127776.9030400@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: Avi Kivity wrote: > Christian Borntr=C3=A4ger wrote: >>> To summarize, Anthony thinks it should use virtio, while I believe >>> virtio is useful for exporting guest memory, not for importing host= =20 >>> memory. >>> =20 >> >> I think the current virtio interface is not ideal for importing host= =20 >> memory, but we can change that. If you look at the dcssblk driver fo= r=20 >> s390, it allows a guest to map shared memory segments via a diagnose= =20 >> (hypercall). This driver uses PCI regions to map memory. >> >> My point is, that the method to map memory is completely irrelevant,= =20 >> we just need something like mmap/shmget between the guest and the=20 >> host. We could define an interface in virtio, that can be used by an= y=20 >> transport. In case of pci this could be a simple pci map operation. >> What do you think about something like: (CCed Rusty) >> =20 >=20 > Exactly. >=20 Agreed.