From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Michael S. Tsirkin" Subject: Re: [PATCH] vfio-pci: Add KVM INTx acceleration Date: Tue, 16 Oct 2012 17:23:08 +0200 Message-ID: <20121016152308.GA14266@redhat.com> References: <20121015202031.23323.72827.stgit@bling.home> <20121016063905.GA8501@redhat.com> <1350395503.2112.209.camel@bling.home> <20121016141443.GA13135@redhat.com> <1350398884.2112.234.camel@bling.home> <20121016150806.GA13732@redhat.com> <1350400395.2112.235.camel@bling.home> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: qemu-devel@nongnu.org, kvm@vger.kernel.org To: Alex Williamson Return-path: Received: from mx1.redhat.com ([209.132.183.28]:33848 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754689Ab2JPPVM (ORCPT ); Tue, 16 Oct 2012 11:21:12 -0400 Content-Disposition: inline In-Reply-To: <1350400395.2112.235.camel@bling.home> Sender: kvm-owner@vger.kernel.org List-ID: On Tue, Oct 16, 2012 at 09:13:15AM -0600, Alex Williamson wrote: > > There's no chance we ship e.g. q35 by mistake without this API: since > > there is no way this specific assert can be missed in even basic > > testing: > > > > So I see it differently: > > > > As coded here: > > chipset authors get lazy and do not implement API. > > bad performance for all users. > > > > With assert: > > chipset authors implement necessary API. > > good performance for all users. > > I prefer a carrot, not a whip. Thanks, > > Alex > It's not just that. Problem is performance testing/fixing is hard. Catching and fixing asserts is easy. So working around buggy qemu code really backfires as it reverses the motivation for writing well performing code. History proves me right: for each API change where we implemented a fallback old code stayed around for years. -- MST From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:44612) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TO8xH-0003X2-Ut for qemu-devel@nongnu.org; Tue, 16 Oct 2012 11:21:25 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1TO8xE-0006DF-0P for qemu-devel@nongnu.org; Tue, 16 Oct 2012 11:21:15 -0400 Received: from mx1.redhat.com ([209.132.183.28]:10953) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TO8xD-0006D2-Nl for qemu-devel@nongnu.org; Tue, 16 Oct 2012 11:21:11 -0400 Date: Tue, 16 Oct 2012 17:23:08 +0200 From: "Michael S. Tsirkin" Message-ID: <20121016152308.GA14266@redhat.com> References: <20121015202031.23323.72827.stgit@bling.home> <20121016063905.GA8501@redhat.com> <1350395503.2112.209.camel@bling.home> <20121016141443.GA13135@redhat.com> <1350398884.2112.234.camel@bling.home> <20121016150806.GA13732@redhat.com> <1350400395.2112.235.camel@bling.home> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1350400395.2112.235.camel@bling.home> Subject: Re: [Qemu-devel] [PATCH] vfio-pci: Add KVM INTx acceleration List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Alex Williamson Cc: qemu-devel@nongnu.org, kvm@vger.kernel.org On Tue, Oct 16, 2012 at 09:13:15AM -0600, Alex Williamson wrote: > > There's no chance we ship e.g. q35 by mistake without this API: since > > there is no way this specific assert can be missed in even basic > > testing: > > > > So I see it differently: > > > > As coded here: > > chipset authors get lazy and do not implement API. > > bad performance for all users. > > > > With assert: > > chipset authors implement necessary API. > > good performance for all users. > > I prefer a carrot, not a whip. Thanks, > > Alex > It's not just that. Problem is performance testing/fixing is hard. Catching and fixing asserts is easy. So working around buggy qemu code really backfires as it reverses the motivation for writing well performing code. History proves me right: for each API change where we implemented a fallback old code stayed around for years. -- MST