From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ian Campbell Subject: Re: [PATCH v9 1/9] xen: vnuma topology and subop hypercalls Date: Fri, 5 Sep 2014 12:45:07 +0100 Message-ID: <1409917507.10156.64.camel@kazak.uk.xensource.com> References: <1409281448-30498-1-git-send-email-ufimtseva@gmail.com> <1409281448-30498-2-git-send-email-ufimtseva@gmail.com> <1409916357.10156.63.camel@kazak.uk.xensource.com> <5409BBB302000078000316C4@mail.emea.novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <5409BBB302000078000316C4@mail.emea.novell.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Jan Beulich Cc: keir@xen.org, lccycc123@gmail.com, george.dunlap@eu.citrix.com, msw@linux.com, dario.faggioli@citrix.com, stefano.stabellini@eu.citrix.com, ian.jackson@eu.citrix.com, xen-devel@lists.xen.org, Elena Ufimtseva List-Id: xen-devel@lists.xenproject.org On Fri, 2014-09-05 at 12:33 +0100, Jan Beulich wrote: > >>> On 05.09.14 at 13:25, wrote: > > On Thu, 2014-08-28 at 23:04 -0400, Elena Ufimtseva wrote: > >> Define interface, structures and hypercalls for toolstack to > >> build vnuma topology and for guests that wish to retrieve it. > >> Two subop hypercalls introduced by patch: > >> XEN_DOMCTL_setvnumainfo to define vNUMA domain topology per domain > >> and XENMEM_get_vnumainfo to retrieve that topology by guest. > > > > This (or a related patch) seems to have broken the build on ARM > > http://www.chiark.greenend.org.uk/~xensrcts/logs/30120/build-armhf/5.ts-xen- > > build.log > > In file included from xenctrl.h:38:0, > > from xc_private.h:32, > > from xc_gnttab.c:20: > > /local/scratch/osstest/osstest/build.30120.build-armhf/xen/tools/libxc/../.. > > /tools/include/xen/domctl.h:1101:33: error: field 'vnuma' has incomplete type > > > > It looks like the definition of xen_domctl_vnuma is under a > > #if defined(__i386__) || defined(__x86_64__) > > but the uses are not. > > > > It would probably be best for now to simply ifdef the uses in libxc too. > > Actually the hypervisor build seems to be broken for ARM too (if you > look a little further down in the log) - we just should move the #endif > up, as there's nothing architecture specific here (even if ARM may > not immediately make use of this). Right, there doesn't seem to be anything x86 specific there. > So Elena, going forward please at least build test your changes on > ARM too. Getting an ARM userspace is a bit tricky, but at least for the hypervisor it cross builds quite easily. I use the cross toolchains from Linaro. Ian.