From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Michael S. Tsirkin" Subject: Re: [PATCH] virtio_mmio: expose header to userspace Date: Fri, 13 Jan 2017 15:31:14 +0200 Message-ID: <20170113152618-mutt-send-email-mst__39060.6218287305$1484314309$gmane$org@kernel.org> References: <1484257033-11032-1-git-send-email-mst@redhat.com> <20170113075822.GA31777@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <20170113075822.GA31777@infradead.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: virtualization-bounces@lists.linux-foundation.org Errors-To: virtualization-bounces@lists.linux-foundation.org To: Christoph Hellwig Cc: linux-kernel@vger.kernel.org, virtualization@lists.linux-foundation.org List-Id: virtualization@lists.linuxfoundation.org On Thu, Jan 12, 2017 at 11:58:22PM -0800, Christoph Hellwig wrote: > On Thu, Jan 12, 2017 at 11:37:22PM +0200, Michael S. Tsirkin wrote: > > It's handy for userspace emulators like QEMU. > > But it's not actually a userspace API - it's an on the write protocol. We do this for all other virtio transports, mmio shouldn't be any different. > so: NAK. We'll have to agree to disagree on this one. It's the interface beween linux and the hypervisor. It isn't any less stable than the one to userspace, for better or worse kvm (ab)uses uapi for that. Compare also to include/uapi/linux/kvm_para.h etc. -- MST