From mboxrd@z Thu Jan 1 00:00:00 1970 From: Juan Quintela Subject: KVM call minutes for 2017-03-14 Date: Mon, 20 Mar 2017 12:15:18 +0100 Message-ID: <8760jbu434.fsf@secure.mitica> Reply-To: quintela@redhat.com Mime-Version: 1.0 Content-Type: text/plain To: QEMU Developer , KVM devel mailing list Return-path: Received: from mx1.redhat.com ([209.132.183.28]:36570 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753846AbdCTLPW (ORCPT ); Mon, 20 Mar 2017 07:15:22 -0400 Sender: kvm-owner@vger.kernel.org List-ID: Hi After a long while, we have had a kvm call. See the thread for the agenda for things that were discussed. My notes are a bit clear because I was also partipating. Please, reply with more things. - What to do with qemu to evolve? Making it more modular (seems to be a good idea) more configurable (decide what we compile in or not) * xen and kvm can already be compiled out * there are patches posted for tcg, but not upstream - how widely use is kvmtool? - Change to a different language? * what we want to achieve wit that? Better support to detect "errors" Less boilerplate code to have to write Easier language for writing emulation code - Can we do anything better than more documentation for people that want to write a driver? - Should we use more upstream libraries? Should we develop more inside qemu? - Object model? * Was QOM a good decision? Better to move to glib model? - Documentation? What format are we going to use There are patches to be integrated. - What can we do to help device model writers? Thanks, Juan.