From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:41360) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1cTpX1-0008Es-Ua for qemu-devel@nongnu.org; Wed, 18 Jan 2017 07:40:05 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1cTpWy-00031z-PG for qemu-devel@nongnu.org; Wed, 18 Jan 2017 07:40:03 -0500 Received: from mx1.redhat.com ([209.132.183.28]:44401) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1cTpWy-00031n-Gr for qemu-devel@nongnu.org; Wed, 18 Jan 2017 07:40:00 -0500 Received: from int-mx14.intmail.prod.int.phx2.redhat.com (int-mx14.intmail.prod.int.phx2.redhat.com [10.5.11.27]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id A0EC114BFFF for ; Wed, 18 Jan 2017 12:40:00 +0000 (UTC) Date: Wed, 18 Jan 2017 10:39:57 -0200 From: Eduardo Habkost Message-ID: <20170118123957.GQ3491@thinpad.lan.raisama.net> References: <20170117010204.4909-1-ehabkost@redhat.com> <20170117010204.4909-10-ehabkost@redhat.com> <9b1bc94e-f1bf-d166-e796-16410a66bb93@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <9b1bc94e-f1bf-d166-e796-16410a66bb93@redhat.com> Subject: Re: [Qemu-devel] [PATCH 9/9] tests: Test case for query-cpu-model-expansion List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: David Hildenbrand Cc: qemu-devel@nongnu.org On Wed, Jan 18, 2017 at 10:39:46AM +0100, David Hildenbrand wrote: > Am 17.01.2017 um 02:02 schrieb Eduardo Habkost: > > + def checkExpansions(self, model, msg): > > + """Perform multiple expansion operations on model, validate results > > + > > + @model is a CpuModelExpansionInfo struct, with some extra keys: > > + * model['runnable'] should be set to True if the CPU model is > > + runnable on this host > > + * model['qom-props'] will be set to the full list of properties for > > + the CPU, if the model is runnable > > + """ > > + exp_s = self.checkOneExpansion(model, 'static', > > + '%s.static' % (msg)) > > + exp_f = self.checkOneExpansion(model, 'full', > > + '%s.full' % (msg)) > > + exp_ss = self.checkOneExpansion(exp_s, 'static', > > + '%s.static.static' % (msg)) > > + exp_sf = self.checkOneExpansion(exp_s, 'full', > > + '%s.static.full' % (msg)) > > + exp_ff = self.checkOneExpansion(exp_f, 'full', > > + '%s.full.full' % (msg)) > > + > > + # static expansion twice should result in the same data: > > + self.assertEquals(exp_s, exp_ss, '%s: static != static+static' % (msg)) > > + # full expansion twice should also result in the same data: > > + self.assertEquals(exp_f, exp_ff, '%s: full != full+full' % (msg)) > > + > > + # migration-safe CPU models have an extra feature: > > + # their static expansion should be equivalent to the full > > + # expansion (as their static expansion is also precise) > > This is not true for s390x: > > "z13-base" is both, static and migration-safe. > > Doing a full expansion will expand all features (so your check against > QOM properties should succeed) > > Doing a static expansion will expand no features, as z13-base is > already static, so there are no features to expand (no delta changes). > > "z13" is only migration-safe. > > Doing a full expansion will expand all features. > > Doing a static expansion will only expand the features different to > "z13-base". (Remember, delta changes only to minimize reported > features). I think my comment was confusing. By "equivalent" I don't mean having the same expansion, but resulting in the same set of features. This is not comparing full_expansion(model) and static_expansion(model). It is comparing full_expansion(model) full_expansion(static_expansion(model)). In other words, absolutely no feature should be lost or changed during static expansion, and we verify that by doing a full expansion after the static expansion (exp_sf) and comparing the results with the full expansion (exp_f). I believe this is true on s390x too, isn't it? > > > And I wonder if that is also true for x86? This should only be true if > the "base" model contains absolutely no features. This is true on x86, but temporarily. "base" still contains no features, but I plan to add extra information to type=full that can't appear on type=static. > > > + if self.isMigrationSafe(model['model']): > > + self.assertEquals(exp_sf['model']['props'], exp_f['model']['props'], > > + '%s: props: static+full != full' % (msg)) > > + self.assertEquals(exp_sf.get('qom-props'), exp_f.get('qom-props'), > > + '%s: qom-props: static+full != full' % (msg)) > > > -- > > David -- Eduardo