From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Is: QXL in Xen (busted) Was :Re: Xen 4.5-rc1 update (RC1 is out 2014-Oct-24th) Date: Tue, 28 Oct 2014 13:18:11 -0400 Message-ID: <20141028171811.GA27336@laptop.dumpdata.com> References: <20141024180843.EA0DF10D709@laptop.dumpdata.com> <544B5AF5.7050103@m2r.biz> <20141027161542.GK4050@laptop.dumpdata.com> <544F987E.5060508@m2r.biz> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mail6.bemta4.messagelabs.com ([85.158.143.247]) by lists.xen.org with esmtp (Exim 4.72) (envelope-from ) id 1XjARo-0004eG-8L for xen-devel@lists.xenproject.org; Tue, 28 Oct 2014 17:20:44 +0000 Content-Disposition: inline In-Reply-To: <544F987E.5060508@m2r.biz> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Fabio Fantoni Cc: artem.mygaiev@globallogic.com, msw@amazon.com, ian.jackson@eu.citrix.com, Steve.VanderLeest@dornerworks.com, mengxu@cis.upenn.edu, m.a.young@durham.ac.uk, chao.p.peng@linux.intel.com, zhigang.x.wang@oracle.com, parth.dixit@linaro.org, boris.ostrovsky@oracle.com, davi.d.vrabel@citrix.com, Paul.Skentzos@dornerworks.com, vijay.kilari@gmail.com, rcojocaru@bitdefender.com, guijianfeng@cn.fujitsu.com, daniel.kiper@oracle.com, josh.whitehead@dornerworks.com, avanzini.arianna@gmail.com, anthony.perard@citrix.com, xen-devel@lists.xenproject.org, serge.broslavsky@linaro.org, feng.wu@intel.com, yjhyun.yoo@samsung.com, olaf@aepfle.de, suriyan.r@gmail.com, ian.campbell@citrix.com, wency@cn.fujitsu.com, stefano.stabellini@eu.citrix.com, mcgrof@suse.com, julien.grall@linaro.org, talex5@gmail.com, robert.vanvossen@dornerworks.com, roy.franz@linaro.org, yang.z.zhang@intel.com, Paul.Dur List-Id: xen-devel@lists.xenproject.org On Tue, Oct 28, 2014 at 02:22:06PM +0100, Fabio Fantoni wrote: > Il 27/10/2014 17:15, Konrad Rzeszutek Wilk ha scritto: > >On Sat, Oct 25, 2014 at 10:10:29AM +0200, Fabio Fantoni wrote: > >>Il 24/10/2014 20:08, konrad.wilk@oracle.com ha scritto: > >>>Feature patchsets that did not make it in by today have been put > >>>on the deferred list. If you think your feature should make it in Xen 4.5-rc2 > >>>please make your case. > >>> > >>>Xen 4.5-rc1 is out today. There are some issues (see 'Known Issues' below) > >>>which are to be fixed in RC2. The official test-day is on Wednesday (Oct 29th) > >>>but if you want to start testing it today - please do! > >>> > >>>Details for the test-day are at > >>> > >>>http://wiki.xen.org/wiki/Xen_4.5_RC1_test_instructions > >>> > >>>In terms of bugs, we have: > >>> > >>>#6 linux: pv drivers miss shutdown command if issued too early > >>>#8 "linux, xenbus mutex hangs when rebooting dom0 and guests hung." > >>>#11 qxl hypervisor support > >>About qxl latest version of libxl/xl support patch is here (also posted to > >>xen-devel) and should be ok: > >>https://github.com/Fantu/Xen/commit/9d9d0ae23ab3e3e1fcfbb6f3201d3a79f958a7a3 > >which is aab4d1b266ce9891a684704f6abf6a5f6b3f7c24 upstrem. > > Was applied upstream for 4.3 and after removed because linux domUs problem > was not solved before stable release. > In github I put also a newer version with some improvements. Ah, so this is a new version that had not yet been posted on xen-devel? It is Xen 4.5 RC1 so I am not too thrilled to put this in the release, especially as it sounds like there are still problems to figure out first. Changing the title to reflect this. > > > > >>With xen 4.5 qxl vga in windows 7 64 bit domUs seems that works good as kvm > >>except for a strange problem with xl save/restore. > >Could you paste in your guest config please. > > name='W7-02' > builder="hvm" > memory=2048 > vcpus=2 > acpi_s3=0 > acpi_s4=0 > vif=['bridge=xenbr0,mac=00:16:3e:42:a2:5f'] > disk=['/mnt/vm/disks/W7-02.disk1.xm,raw,hda,rw',',raw,hdb,ro,cdrom'] > boot='dc' > device_model_version="qemu-xen" > viridian=1 > vnc=0 > keymap="it" > on_crash="destroy" > vga="qxl" > spice=1 > spicehost='0.0.0.0' > spiceport=6003 > spicedisable_ticketing=1 > spicevdagent=1 > spice_clipboard_sharing=0 > spiceusbredirection=4 Does SPICE always have to be enabled to use QXL? > soundhw="hda" > localtime=1 > usbversion=2 > > winpv installed now and on latest tests: > http://fantu.info/xen/winpv-20141020-free.7z > tried also different versions of james harper's pv with same result > and spice guest tools 0.74 (for vdagent and qxl driver) > > > > >>After restore connecting with spice client domU "freeze" for 2-3 minutes, no > >>errors or warnings visible in logs, I tried many times to take more > >>informations about the problem and I already posted about it but for now > >>nothing :( > >>The only probably useful data are that domU's qemu process when the problem > >>happen seems still working but domU's timing (from xl list) not increase. > >>After restore seems do also other strange thing, "freeze" for 2-3 minutes > >>also on windows poweroff. > >That sounds like a timing issue. As in something is waiting for something > >and we are not triggering it. > > > >>Another qxl problem is with linux domUs, I tried to find useful data to > >>solves it for years, I also did many posts about it in xen-devel- qemu-devel > >>and spice-devel. > >>On latest tests domU crash (not qemu crash, I already tried qemu backtrace > >>but seems impossible) on domU's xorg start (with qxl driver installed) and > >>I'm unable to take useful informations in logs and various debugs tried. > >OK, so Linux is not working at all then? > > Before "x86_emulate: properly do IP updates and other side effects on > success" patch always black screen and xorg at 100% cpu. > After that commit domU crash after xorg start with qxl driver installed and > I'm unable to take backtrace or useful data about it. OK, looking at how the driver is setup - it seems that there is an QXL KMS driver. If you boot without Xorg running but just with in the console can you use 'fbset' or any other framebuffer tools (linux-fb-tools)? That is I am curious to see whether the 'FB' is working properly first before figuring out whether Xorg is working.