- Devel - libvirt List Archives

[libvirt] [PATCH] spec file enhancements for minimal "client-only" build
by Ben Guthro 17 Sep '08

17 Sep '08

Attached is a patch to add some additional options to the spec file, allowing for more flexibility when building. With this patch, it allows you to build a "client-only" version of libvirt for machines that may have an interest in communicating over the remote driver, but do not have a hypervisor themselves. It enables command line overrides of control variables for the build eg: rpmbuild $(RPM_FLAGS) --nodeps \ --without xen \ --without xen_proxy \ --without qemu \ --without openvz \ --without lxc \ --without polkit \ --without libvirtd \ --without avahi \ --without sasl \ --without python \ --without devel \ -ta --clean libvirt*.tar.gz Signed-off-by: Ben Guthro <bguthro(a)virtualiron.com> diff --git a/libvirt.spec.in b/libvirt.spec.in index 18b2127..666fb83 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1,11 +1,16 @@ # -*- rpm-spec -*- -%define with_xen 1 -%define with_xen_proxy 1 -%define with_qemu 1 -%define with_openvz 1 -%define with_lxc 1 -%define with_polkit 0 +%define with_xen 0%{!?_without_xen:1} +%define with_xen_proxy 0%{!?_without_xen_proxy:1} +%define with_qemu 0%{!?_without_qemu:1} +%define with_openvz 0%{!?_without_openvz:1} +%define with_lxc 0%{!?_without_lxc:1} +%define with_sasl 0%{!?_without_sasl:1} +%define with_avahi 0%{!?_without_avahi:1} +%define with_polkit 0%{!?_without_polkit:0} +%define with_python 0%{!?_without_python:1} +%define with_libvirtd 0%{!?_without_libvirtd:1} +%define with_devel 0%{!?_without_devel:1} # Xen is available only on i386 x86_64 ia64 %ifnarch i386 i686 x86_64 ia64 @@ -23,7 +28,7 @@ %endif %if 0%{fedora} >= 8 -%define with_polkit 1 +%define with_polkit 0%{!?_without_polkit:1} %define with_xen_proxy 0 %endif @@ -46,10 +51,12 @@ Requires: iptables # So remote clients can access libvirt over SSH tunnel # (client invokes 'nc' against the UNIX socket on the server) Requires: nc +%if %{with_sasl} Requires: cyrus-sasl # Not technically required, but makes 'out-of-box' config # work correctly & doesn't have onerous dependencies Requires: cyrus-sasl-md5 +%endif %if %{with_polkit} Requires: PolicyKit >= 0.6 %endif @@ -82,14 +89,18 @@ BuildRequires: readline-devel BuildRequires: ncurses-devel BuildRequires: gettext BuildRequires: gnutls-devel +%if %{with_avahi} BuildRequires: avahi-devel +%endif BuildRequires: libselinux-devel BuildRequires: dnsmasq BuildRequires: bridge-utils %if %{with_qemu} BuildRequires: qemu %endif +%if %{with_sasl} BuildRequires: cyrus-sasl-devel +%endif %if %{with_polkit} BuildRequires: PolicyKit-devel >= 0.6 %endif @@ -119,6 +130,7 @@ BuildRequires: gawk Libvirt is a C toolkit to interact with the virtualization capabilities of recent versions of Linux (and other OSes). +%if %{with_devel} %package devel Summary: Libraries, includes, etc. to compile with the libvirt library Group: Development/Libraries @@ -132,7 +144,9 @@ Obsoletes: libvir-devel %description devel Includes and documentations for the C library providing an API to use the virtualization capabilities of recent versions of Linux (and other OSes). +%endif +%if %{with_python} %package python Summary: Python bindings for the libvirt library Group: Development/Libraries @@ -144,6 +158,7 @@ The libvirt-python package contains a module that permits applications written in the Python programming language to use the interface supplied by the libvirt library to use the virtualization capabilities of recent versions of Linux (and other OSes). +%endif %prep %setup -q @@ -190,6 +205,7 @@ install -d -m 0755 $RPM_BUILD_ROOT%{_localstatedir}/lib/libvirt/images/ # Default dir for kernel+initrd images defnied in SELinux policy install -d -m 0755 $RPM_BUILD_ROOT%{_localstatedir}/lib/libvirt/boot/ +%if %{with_qemu} # We don't want to install /etc/libvirt/qemu/networks in the main %files list # because if the admin wants to delete the default network completely, we don't # want to end up re-incarnating it on every RPM upgrade. @@ -200,14 +216,33 @@ rm -f $RPM_BUILD_ROOT%{_sysconfdir}/libvirt/qemu/networks/default.xml rm -f $RPM_BUILD_ROOT%{_sysconfdir}/libvirt/qemu/networks/autostart/default.xml # Strip auto-generated UUID - we need it generated per-install sed -i -e "/<uuid>/d" $RPM_BUILD_ROOT%{_datadir}/libvirt/networks/default.xml +%endif %find_lang %{name} +# If we are not building with certain options, we need to remove files +# so that we don't get Installed (but unpackaged) errors +%if ! %{with_devel} +rm -rf $RPM_BUILD_ROOT%{_datadir}/gtk-doc/html/libvirt +rm -rf $RPM_BUILD_ROOT%{_includedir}/libvirt +rm -rf $RPM_BUILD_ROOT%{_libdir}/lib*.so +rm -rf $RPM_BUILD_ROOT%{_libdir}/pkgconfig/libvirt.pc +%endif + +%if ! %{with_python} +rm -rf $RPM_BUILD_ROOT%{_datadir}/doc/libvirt-python-%{version} +%endif + +%if ! %{with_qemu} +rm -rf $RPM_BUILD_ROOT%{_sysconfdir}/libvirt/qemu.conf +%endif + %clean rm -fr %{buildroot} %post /sbin/ldconfig +%if %{with_libvirtd} # We want to install the default network for initial RPM installs # or on the first upgrade from a non-network aware libvirt only. # We check this by looking to see if the daemon is already installed @@ -222,12 +257,15 @@ then fi /sbin/chkconfig --add libvirtd +%endif %preun +%if %{with_libvirtd} if [ $1 = 0 ]; then /sbin/service libvirtd stop 1>/dev/null 2>&1 /sbin/chkconfig --del libvirtd fi +%endif %postun /sbin/ldconfig @@ -240,40 +278,73 @@ fi %{_bindir}/virsh %{_libdir}/lib*.so.* %dir %attr(0700, root, root) %{_sysconfdir}/libvirt/ + +%if %{with_qemu} %dir %attr(0700, root, root) %{_sysconfdir}/libvirt/qemu/ %dir %attr(0700, root, root) %{_sysconfdir}/libvirt/qemu/networks/ %dir %attr(0700, root, root) %{_sysconfdir}/libvirt/qemu/networks/autostart +%endif + +%if %{with_libvirtd} %{_sysconfdir}/rc.d/init.d/libvirtd %config(noreplace) %{_sysconfdir}/sysconfig/libvirtd %config(noreplace) %{_sysconfdir}/libvirt/libvirtd.conf +%endif + +%if %{with_qemu} %config(noreplace) %{_sysconfdir}/libvirt/qemu.conf +%endif + +%if %{with_sasl} %config(noreplace) %{_sysconfdir}/sasl2/libvirt.conf +%endif + +%if %{with_qemu} %dir %{_datadir}/libvirt/ %dir %{_datadir}/libvirt/networks/ %{_datadir}/libvirt/networks/default.xml +%endif + %dir %{_localstatedir}/run/libvirt/ %dir %{_localstatedir}/lib/libvirt/ %dir %attr(0700, root, root) %{_localstatedir}/lib/libvirt/images/ %dir %attr(0700, root, root) %{_localstatedir}/lib/libvirt/boot/ -%{_datadir}/augeas/lenses/libvirtd.aug + +%if %{with_qemu} %{_datadir}/augeas/lenses/libvirtd_qemu.aug -%{_datadir}/augeas/lenses/tests/test_libvirtd.aug %{_datadir}/augeas/lenses/tests/test_libvirtd_qemu.aug +%endif + +%if %{with_libvirtd} +%{_datadir}/augeas/lenses/libvirtd.aug +%{_datadir}/augeas/lenses/tests/test_libvirtd.aug +%endif + %if %{with_polkit} %{_datadir}/PolicyKit/policy/org.libvirt.unix.policy %endif + +%if %{with_qemu} %dir %attr(0700, root, root) %{_localstatedir}/log/libvirt/qemu/ +%endif + %if %{with_xen_proxy} %attr(4755, root, root) %{_libexecdir}/libvirt_proxy %endif -%attr(0755, root, root) %{_libexecdir}/libvirt_parthelper + %if %{with_lxc} %attr(0755, root, root) %{_libexecdir}/libvirt_lxc %endif + +%if %{with_libvirtd} +%attr(0755, root, root) %{_libexecdir}/libvirt_parthelper %attr(0755, root, root) %{_sbindir}/libvirtd +%endif + %doc docs/*.rng %doc docs/*.xml +%if %{with_devel} %files devel %defattr(-, root, root) @@ -289,7 +360,9 @@ fi %doc docs/*.html docs/html docs/*.gif %doc docs/examples %doc docs/libvirt-api.xml +%endif +%if %{with_python} %files python %defattr(-, root, root) @@ -300,6 +373,7 @@ fi %doc python/TODO %doc python/libvirtclass.txt %doc docs/examples/python +%endif %changelog * Mon Sep 8 2008 Daniel Veillard <veillard(a)redhat.com> - 0.4.5-1

3 10

[libvirt] [PATCH] kvm: maxVCPU runtime detection
by Guido Günther 17 Sep '08

17 Sep '08

Hi, with recent linux kernels we can detect the maximum number of virtual cpus at runtime via an ioctl. Possible patch attached. It does this on every call to qemudGetMaxVCPUs. Would you prefer something that does this only once in qemudStartup()? Cheers, -- Guido

7 24

[libvirt] Calling libvirt from C# (Mono/.Net)
by Richard W.M. Jones 17 Sep '08

17 Sep '08

As an experiment I managed to call libvirt from C# (using Mono, but similar code should work for .Net). This email should give interested people enough information to produce a proper set of bindings for libvirt. ---------------------------------------------------------------------- using System; using System.Runtime.InteropServices; [StructLayout (LayoutKind.Sequential)] class NodeInfo { [MarshalAs (UnmanagedType.ByValTStr, SizeConst=32)] public string model; public ulong memory; public uint cpus; public uint mhz; public uint nodes; public uint sockets; public uint cores; public uint threads; } class Connection { IntPtr _handle; [DllImport ("libvirt.so")] static extern IntPtr virConnectOpen ([In] string name); public Connection (string name) { _handle = virConnectOpen (name); } [DllImport ("libvirt.so")] static extern IntPtr virConnectClose (IntPtr h); ~Connection () { virConnectClose (_handle); } [DllImport ("libvirt.so")] static extern int virNodeGetInfo (IntPtr h, [Out] NodeInfo info); public NodeInfo nodeGetInfo () { NodeInfo info = new NodeInfo (); if (virNodeGetInfo (_handle, info) == -1) { // throw new LibvirtException ("virNodeGetInfo"); } return info; } } class Test { public static void Main () { Connection conn = new Connection ("test:///default"); NodeInfo info = conn.nodeGetInfo (); Console.WriteLine ("Results of virNodeGetInfo:"); Console.WriteLine ("model = {0}", info.model); Console.WriteLine ("memory = {0}", info.memory); Console.WriteLine ("cpus = {0}", info.cpus); Console.WriteLine ("mhz = {0}", info.mhz); Console.WriteLine ("nodes = {0}", info.nodes); Console.WriteLine ("sockets = {0}", info.sockets); Console.WriteLine ("cores = {0}", info.cores); Console.WriteLine ("threads = {0}", info.threads); } } ---------------------------------------------------------------------- Compile and run this code with: $ gmcs libvirt-test.c# $ LIBVIRT_DEBUG=1 mono ./libvirt-test.exe ---------------------------------------------------------------------- DEBUG: libvirt.c: virInitialize (register drivers) DEBUG: libvirt.c: virConnectOpen (name=test:///default) DEBUG: libvirt.c: do_open (name "test:///default" to URI components: scheme test opaque (null) authority (null) server (null) user (null) port 0 path /default ) DEBUG: libvirt.c: do_open (trying driver 0 (Test) ...) DEBUG: libvirt.c: do_open (driver 0 Test returned SUCCESS) DEBUG: libvirt.c: do_open (network driver 0 Test returned SUCCESS) DEBUG: libvirt.c: do_open (storage driver 0 Test returned SUCCESS) DEBUG: libvirt.c: virNodeGetInfo (conn=0xa8d560, info=0x7fff9530c070) Results of virNodeGetInfo: model = i686 memory = 3145728 cpus = 16 mhz = 1400 nodes = 2 sockets = 2 cores = 2 threads = 2 DEBUG: libvirt.c: virConnectClose (conn=0xa8d560) DEBUG: hash.c: virUnrefConnect (unref connection 0xa8d560 test:///default 1) DEBUG: hash.c: virReleaseConnect (release connection 0xa8d560 test:///default) ---------------------------------------------------------------------- Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://et.redhat.com/~rjones/virt-top

1 0

[libvirt] network domain name patch
by JJ Reynolds 16 Sep '08

16 Sep '08

Hi all, This is a followup to the excellent patch which allows static IP address assignment. The problem with that patch is that you can only set the host name and cannot set a FQDN because dnsmasq will, as a security measure, not allow it unless --domain is specified. This patch adds support for adding <domain name="my.domain" /> to the network config file. With that stanza, one can then use FQDNs on the static host assignments, and this should be the domain reported for any clients that request it. If <domain name .. ./> is not specified in the config file, then there is no change in behaviour. As a special case, you can also set the domain name to "#", whereupon dnsmasq interprets that to use the domain of the host OS. Example default.xml: <network> <name>default</name> <uuid>0098abb7-ff94-4df9-aa78-e4c3fe636a3d</uuid> <bridge name="virbr0" /> <domain name="mynet.net" /> <forward/> <ip address="192.168.122.1" netmask="255.255.255.0"> <dhcp> <host name="vm1.mynet.net" mac="00:16:3e:24:a5:84" ip="192.168.122.101" /> <host name='vm2.mynet.net' mac="00:16:3e:1f:9a:95" ip="192.168.122.102" /> <range start="192.168.122.2" end="192.168.122.99" /> </dhcp> </ip> </network> Here is the patch: diff -r -U 3 libvirt-0.4.4/src/network_conf.c libvirt-0.4.4jjr/src/network_conf.c --- libvirt-0.4.4/src/network_conf.c 2008-08-20 09:20:01.000000000 -0400 +++ libvirt-0.4.4jjr/src/network_conf.c 2008-09-06 14:03:56.000000000 -0400 @@ -326,6 +326,9 @@ VIR_FREE(tmp); } + /* Parse network domain information */ + def->domain = virXPathString(conn, "string(./domain[1]/@name)", ctxt); + /* Parse bridge information */ def->bridge = virXPathString(conn, "string(./bridge[1]/@name)", ctxt); tmp = virXPathString(conn, "string(./bridge[1]/@stp)", ctxt); diff -r -U 3 libvirt-0.4.4/src/network_conf.h libvirt-0.4.4jjr/src/network_conf.h --- libvirt-0.4.4/src/network_conf.h 2008-08-20 09:20:01.000000000 -0400 +++ libvirt-0.4.4jjr/src/network_conf.h 2008-09-06 14:03:56.000000000 -0400 @@ -57,6 +57,7 @@ char *name; char *bridge; /* Name of bridge device */ + char *domain; unsigned long delay; /* Bridge forward delay (ms) */ int stp : 1; /* Spanning tree protocol */ diff -r -U 3 libvirt-0.4.4/src/qemu_driver.c libvirt-0.4.4jjr/src/qemu_driver.c --- libvirt-0.4.4/src/qemu_driver.c 2008-08-29 03:20:02.000000000 -0400 +++ libvirt-0.4.4jjr/src/qemu_driver.c 2008-09-06 14:15:28.000000000 -0400 @@ -1105,6 +1105,7 @@ 1 + /* --keep-in-foreground */ 1 + /* --strict-order */ 1 + /* --bind-interfaces */ + (network->def->domain?2:0) + /* --domain name */ 2 + /* --pid-file "" */ 2 + /* --conf-file "" */ /*2 + *//* --interface virbr0 */ @@ -1136,6 +1137,11 @@ APPEND_ARG(*argv, i++, "--strict-order"); APPEND_ARG(*argv, i++, "--bind-interfaces"); + if (network->def->domain) { + APPEND_ARG(*argv, i++, "--domain"); + APPEND_ARG(*argv, i++, network->def->domain); + } + APPEND_ARG(*argv, i++, "--pid-file"); APPEND_ARG(*argv, i++, ""); Thanks! -JJ Reynolds

3 11

[libvirt] LXC: making the private root filesystem more secure
by Daniel P. Berrange 12 Sep '08

12 Sep '08

When I wrote the private root filesystem stuff for LXC (which I just committed) I noted that we couldn't actually make this secure, because someone inside the chroot can just 'mknod' and access the host devices. What I completely forgot was that cgroups as of 2.6.26 has device ACLs If we place every container in a cgroup (which was planned anyway), then we can trivially prevent containers accessing host devices One time setup mount -t cgroups /dev/cgroups mkdir /dev/cgroups/libvirt mkdir /dev/cgroups/libvirt/lxc For each new container 'NAME' mkdir /dev/cgroups/libvirt/lxc/{NAME} echo "a" > /dev/cgroups/libvirt/lxc/{NAME}/devices.deny echo "c 1:3 rwm" > /dev/cgroups/libvirt/lxc/{NAME}/devices.allow echo "c 1:5 rwm" > /dev/cgroups/libvirt/lxc/{NAME}/devices.allow echo "c 1:7 rwm" > /dev/cgroups/libvirt/lxc/{NAME}/devices.allow echo "c 5:1 rwm" > /dev/cgroups/libvirt/lxc/{NAME}/devices.allow echo "c 1:8 rwm" > /dev/cgroups/libvirt/lxc/{NAME}/devices.allow echo "c 1:9 rwm" > /dev/cgroups/libvirt/lxc/{NAME}/devices.allow This denies all devices, and then allows null, zero, full, console, random and urandom. Allowing use of 'random' is debatable. The 'devpts' namespace stuff is also needed to provide private PTYs. The 'user' namespace stuff is needed to prevent an unprivileged user in the host OS from killing off processes with same UID inside the container. There looks to be active patchsets for both of these being discussed, so we're getting close to having a genuinely useful container based virt driver with LXC Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|

5 11

[libvirt] virDomainGetInfo() returns 0 although the domain is not running anymore
by Johannes Formann 12 Sep '08

12 Sep '08

Hello, if I understand the API-documentation, virDomainGetInfo() should return -1 when the domain does not exists anymore, or some other problem happens. But with this code below ... if (virDomainGetInfo(domainPtr,domainInfo) == 0 ) { if (debugLevel > 15) {puts("virDomainGetInfo() returns 0");} if (domainInfo->state == VIR_DOMAIN_NOSTATE || domainInfo->state == VIR_DOMAIN_RUNNING || domainInfo->state == VIR_DOMAIN_BLOCKED || domainInfo->state == VIR_DOMAIN_PAUSED ) { return(false); } } ... I notized, virDomainGetInfo() always returns zero, even if libvirt prints an error. If I destroy a domain (xm destroy) I'll get this Message with each call of virDomainGetInfo(): "libvir: Xen Daemon error : GET operation failed:" But virDomainGetInfo() still returns zero, a bug, or did I missunderstood the API? regards Johannes

2 2

[libvirt] qemu or kvm?
by Jun Koi 12 Sep '08

12 Sep '08

Hi, Given a virDomainPtr variable, how can we know that it is a QEMU or KVM machine? Is there any API for that? Many thanks, Jun

2 3

[libvirt] [PATCH] Fix segfault if no qemu emulator passed
by Cole Robinson 12 Sep '08

12 Sep '08

There is unfortunately a pretty prevalent segfault in the latest libvirt. If a qemu domain is defined without an emulator specified, libvirtd crashes. This is doubly unfortunate since current virtinst generates an emulator-less config for all qemu and kvm guests (I'm about to fix this upstream though). Attached patch fixes this. Thanks, Cole diff --git a/src/qemu_conf.c b/src/qemu_conf.c index d742c32..23ef050 100644 --- a/src/qemu_conf.c +++ b/src/qemu_conf.c @@ -790,7 +790,7 @@ int qemudBuildCommandLine(virConnectPtr conn, if (!emulator) return -1; - ADD_ARG_LIT(vm->def->emulator); + ADD_ARG_LIT(emulator); ADD_ARG_LIT("-S"); ADD_ARG_LIT("-M"); ADD_ARG_LIT(vm->def->os.machine);

6 16

[libvirt] Sharing file between host and guest
by Yushu Yao 12 Sep '08

12 Sep '08

Hi All, Question not very related to libvirt but I think most of the real VM experts are in this list. What is the best and fastest way to share files between guest and host? Something like the share folders-hgfs in vmware might work, but is there a way that works on all hypervisors? Network based file sharing (NFS/Samba) are limited by the speed of the virtual network interfaces. Maybe one can do it through USB? Thanks a lot! -yushu

2 1

[libvirt] [PATCH] make consistent assumptions about libvirtd dependence
by David Lively 12 Sep '08

12 Sep '08

Hi - Currently src/Makefile.am says that only the QEMU, LXC, and storage drivers are dependent on libvirtd. But the driver registration code in virInitialize wasn't registering the remote or openVz driver when WITH_LIBVIRTD was not defined. The attached patch makes the virInitialize code consistent with the assumptions stated in src/Makefile.am. But note I'm assuming the assumptions in src/Makefile.am are correct. I'm not sure what's really dependent on libvirtd, so Someone Who Knows Better should verify this. Thanks, Dave commit 24d5708877fce90c4e85e3b2ede07b74f4a98224 Author: David Lively <dlively(a)virtualiron.com> Date: Thu Sep 11 11:56:20 2008 -0400 Make driver registration assume the same libvirtd dependences expressed in src/Makefile.am (i.e., that only the QEMU, LXC, and storage drivers depend on libvirtd). diff --git a/src/libvirt.c b/src/libvirt.c index 54ed8cf..3d66f5a 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -286,20 +286,20 @@ virInitialize(void) #ifdef WITH_XEN if (xenUnifiedRegister () == -1) return -1; #endif -#ifdef WITH_LIBVIRTD -#ifdef WITH_QEMU - if (qemudRegister() == -1) return -1; +#ifdef WITH_REMOTE + if (remoteRegister () == -1) return -1; #endif #ifdef WITH_OPENVZ if (openvzRegister() == -1) return -1; #endif +#ifdef WITH_LIBVIRTD +#ifdef WITH_QEMU + if (qemudRegister() == -1) return -1; +#endif #ifdef WITH_LXC if (lxcRegister() == -1) return -1; #endif if (storageRegister() == -1) return -1; -#ifdef WITH_REMOTE - if (remoteRegister () == -1) return -1; -#endif #endif return(0);

2 1