From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:60316) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XYYVR-0007Op-KE for qemu-devel@nongnu.org; Mon, 29 Sep 2014 06:48:43 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XYYVL-0001sE-MG for qemu-devel@nongnu.org; Mon, 29 Sep 2014 06:48:37 -0400 Received: from david.siemens.de ([192.35.17.14]:41601) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XYYVL-0001kh-AH for qemu-devel@nongnu.org; Mon, 29 Sep 2014 06:48:31 -0400 Message-ID: <542938EF.3040207@siemens.com> Date: Mon, 29 Sep 2014 12:48:15 +0200 From: Jan Kiszka MIME-Version: 1.0 References: <1411718914-6608-1-git-send-email-edgar.iglesias@gmail.com> <542934E4.3010902@siemens.com> In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH v7 00/11] target-arm: Parts of the AArch64 EL2/3 exception model List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Peter Maydell Cc: Rob Herring , Peter Crosthwaite , Fabian Aggeler , Greg Bellows , QEMU Developers , Alexander Graf , Paolo Bonzini , Sergey Fedorov , "Edgar E. Iglesias" , =?UTF-8?B?QWxleCBCZW5uw6ll?= , Christoffer Dall , Richard Henderson On 2014-09-29 12:41, Peter Maydell wrote: > On 29 September 2014 11:31, Jan Kiszka wrote: >> Sorry for hijacking the thread, but it seems related: These bits address >> AArch64, but what is the status of AArch32 /wrt hyp mode emulation? >> After playing with the "fast" model, I would be glad to find such >> support in QEMU rather sooner than later. > > I don't think anybody is currently working on that, > though it is an obvious gap to fill in at some point. > > I would caution against testing hypervisors on QEMU > rather than against hardware or the Fast Model, > except for anything other than basic smoke tests, > though -- there are a lot of complications and corner > cases in the architecture here and I expect we will > spend some time flushing out bugs as OSes expose > them. No difference with nested virt for KVM: I fixed several corner cases along the way of getting Jailhouse running on x86. This time we already have a working baseline, so shaking out bugs might be even easier. What matters more mid- and long-term is the round-trip time, in fact. Jan -- Siemens AG, Corporate Technology, CT RTC ITP SES-DE Corporate Competence Center Embedded Linux