From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:57752) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1flWP5-0005d9-Rl for qemu-devel@nongnu.org; Fri, 03 Aug 2018 05:29:48 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1flWP0-0005qR-Tu for qemu-devel@nongnu.org; Fri, 03 Aug 2018 05:29:47 -0400 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:52670 helo=mx1.redhat.com) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1flWP0-0005o6-Q0 for qemu-devel@nongnu.org; Fri, 03 Aug 2018 05:29:42 -0400 Date: Fri, 3 Aug 2018 17:29:39 +0800 From: Fam Zheng Message-ID: <20180803092939.GF16395@lemon.usersys.redhat.com> References: <20180803085230.30574-1-peter.maydell@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180803085230.30574-1-peter.maydell@linaro.org> Subject: Re: [Qemu-devel] [PATCH 0/5] tests/vm: various minor improvements List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Peter Maydell Cc: qemu-devel@nongnu.org, Alex =?iso-8859-1?Q?Benn=E9e?= , Philippe =?iso-8859-1?Q?Mathieu-Daud=E9?= , patches@linaro.org On Fri, 08/03 09:52, Peter Maydell wrote: > The setup I had for my BSD VMs broke today, so I thought I'd > have a look at using the tests/vm/ support for building QEMU > inside BSD VMs rather than re-rolling my own. This patchset > fixes some rough edges I ran into along the way: > * fixes support for non-KVM accelerators > * propagates J=n setting into 'make check' in the VM > * propagates V=1 setting into 'make' and 'make check' in the VM > * uses --output-sync so that J=n settings don't mangle output > * bumps the RAM settings so J=n doesn't cause the guest to > run out of memory and kill the compiler > > Ideally I'd like to be able to separately invoke commands > for "build in VM" and "run make check in VM for that build"; for > instance I only want V=1 on the make-check, not the make, and in > some cases might want to do a build but not check, and so on. > But that seemed too complicated for the moment. If we can limit the abstraction to "build" and "check" (or "test"), it shouldn't be hard to add a --build-only option to the Python main() and VM_BUILD_ONLY=1 to Makefile. > > Is it possible for the VMs to be persistent? Currently it > looks like they throw away their contents and start afresh > for every invocation. Doing build tests as incremental is > faster than doing every one as a from-clean... Yes, we'll do that in 3.1. > > Also, I notice that the OpenBSD VM setup is using SDL 1.2, > which means that configure complains and suggests switching > to SDL 2.0. How do we update the base image? I prepared the image by hand and uploaded it because I couldn't find a better way. So the easist way to update is probably booting the image by hand, do the modifications, then upload it again. > > thanks > -- PMM > > Peter Maydell (5): > tests/vm: Use -cpu max rather than -cpu host > tests/vm: Pass the jobs parallelism setting to 'make check' > tests/vm: Propagate V=1 down into the make inside the VM > tests/vm: Bump guest RAM up from 2G to 4G > tests/vm: Use make's --output-sync option > > tests/vm/Makefile.include | 1 + > docs/devel/testing.rst | 1 + > tests/vm/basevm.py | 9 ++++++--- > tests/vm/freebsd | 4 ++-- > tests/vm/netbsd | 4 ++-- > tests/vm/openbsd | 4 ++-- > tests/vm/ubuntu.i386 | 4 ++-- > 7 files changed, 16 insertions(+), 11 deletions(-) > > -- > 2.17.1 > >