- Devel - libvirt List Archives

[libvirt] [PATCH 0/5] Various logging cleanups
by Amy Griffis 06 Aug '09

06 Aug '09

The following patches fixup libvirt's logging code, to hopefully make it a little more consistent and predictable. --- Amy Griffis (5): Tighten libvirt's parsing of logging environment variables Several fixes to libvirtd's log setup Cleanup VIR_LOG_DEBUG parsing in eventtest Consolidate code for parsing the logging environment variables Update logging documentation docs/logging.html.in | 22 +++++++- qemud/qemud.c | 90 +++++++++++++++++--------------- src/libvirt.c | 19 ------- src/libvirt_private.syms | 5 ++ src/logging.c | 129 +++++++++++++++++++++++++++++++++++++++------- src/logging.h | 5 ++ tests/eventtest.c | 12 +--- 7 files changed, 190 insertions(+), 92 deletions(-)

4 19

[libvirt] [PATCH] chown kernel/initrd before spawning qemu
by Mark McLoughlin 06 Aug '09

06 Aug '09

If we're running qemu unprivileged, we need to chown any supplied kernel or initrd before spawning it. * src/qemu_driver.c: rename qemuDomainSetDiskOwnership() to qemuDomainSetFileOwnership(), pass it a path string instead of a disk definition and use it for chowning the kernel/initrd in qemuDomainSetAllDeviceOwnership() --- src/qemu_driver.c | 20 ++++++++++++-------- 1 files changed, 12 insertions(+), 8 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 412b68d..bd58435 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -1684,18 +1684,18 @@ static int qemuDomainSetHostdevOwnership(virConnectPtr conn, } -static int qemuDomainSetDiskOwnership(virConnectPtr conn, - virDomainDiskDefPtr def, +static int qemuDomainSetFileOwnership(virConnectPtr conn, + const char *path, uid_t uid, gid_t gid) { - if (!def->src) + if (!path) return 0; - VIR_DEBUG("Setting ownership on %s to %d:%d", def->src, uid, gid); - if (chown(def->src, uid, gid) < 0) { + VIR_DEBUG("Setting ownership on %s to %d:%d", path, uid, gid); + if (chown(path, uid, gid) < 0) { virReportSystemError(conn, errno, _("cannot set ownership on %s"), - def->src); + path); return -1; } return 0; @@ -1725,7 +1725,7 @@ static int qemuDomainSetDeviceOwnership(virConnectPtr conn, (def->data.disk->readonly || def->data.disk->shared)) return 0; - return qemuDomainSetDiskOwnership(conn, def->data.disk, uid, gid); + return qemuDomainSetFileOwnership(conn, def->data.disk->src, uid, gid); case VIR_DOMAIN_DEVICE_HOSTDEV: return qemuDomainSetHostdevOwnership(conn, def->data.hostdev, uid, gid); @@ -1753,12 +1753,16 @@ static int qemuDomainSetAllDeviceOwnership(virConnectPtr conn, uid = restore ? 0 : driver->user; gid = restore ? 0 : driver->group; + if (qemuDomainSetFileOwnership(conn, def->os.kernel, uid, gid) < 0 || + qemuDomainSetFileOwnership(conn, def->os.initrd, uid, gid) < 0) + return -1; + for (i = 0 ; i < def->ndisks ; i++) { if (restore && (def->disks[i]->readonly || def->disks[i]->shared)) continue; - if (qemuDomainSetDiskOwnership(conn, def->disks[i], uid, gid) < 0) + if (qemuDomainSetFileOwnership(conn, def->disks[i]->src, uid, gid) < 0) return -1; } -- 1.6.2.5

2 3

[libvirt] VMware ESX 3/3: Cleanup varaible naming
by Matthias Bolte 06 Aug '09

06 Aug '09

Subject says it all. No functional change included. Matthias

3 2

[libvirt] virt-manager
by Michael Zeller 06 Aug '09

06 Aug '09

okay so using virt-manager it only lists my root partition as storage, or atleast thats what it appears So it says I can only make a vm the size of 9G when my home partition has 800+G is there a way to change that?

2 1

[libvirt] libvirt-0.3.3-14.el5 (redhat rpm) vs libvirt-0.6.5 (source)
by Jim Bergensen 06 Aug '09

06 Aug '09

Hello everyone, I am looking to write a simple php interface to create, stop, start, restart, configure, backup, transfer VM's across a network of servers. Storage is done locally on the particular node (lvm), but I am considering to do this through iscsi in the future. Xen, KVM, IP management, firewalling is an important requirement. The redhat backport version looks outdated, but I know the version number isn't the best way to compare features. My questions are: - Featurewise, is the difference between the redhat version and the libvirt.org source version small or as big as it looks (branch versions?) - Considering the above functions I am going to need, will the redhat version suffice? Any help would be truely appreciated, Thank you, Jim

4 4

[Fwd: Re: [libvirt] Release of libvirt-0.7.0]
by Robert L Cochran 06 Aug '09

06 Aug '09

Ooops! I sent this to Dan when I meant to send it to the list. To be a bit repetitive...how do I get this for Fedora 11 -- and all of its dependencies? Thanks Bob -------- Original Message -------- Subject: Re: [libvirt] Release of libvirt-0.7.0 Date: Wed, 05 Aug 2009 16:37:09 -0400 From: Robert L Cochran <cochranb(a)speakeasy.net> To: veillard(a)redhat.com How do I get this for Fedora 11 -- and all of its dependencies? Thanks a lot! Bob On 08/05/2009 11:15 AM, Daniel Veillard wrote: > A couple days later than expected, but considering the current > flow of fixes, that's not a bad thing. This is a huge release, > this includes more than 250 commits in a month and many new > functionalities or drivers, and a lot of improvements and bug fixes: > > New features: > - Interface implementation based on netcf (Laine Stump Daniel Veillard) > - Add new net filesystem glusterfs (Harshavardhana) > - Initial VMWare ESX driver (Matthias Bolte) > - Add support for VBox 3 and event callbacks on vbox (Pritesh Kothari) > - First version of the Power Hypervisor driver (Eduardo Otubo) > - Run QEMU guests as an unprivileged user (Daniel P. Berrange) > - Support cgroups in QEMU driver (Daniel P. Berrange) > - QEmu hotplug NIC support (Mark McLoughlin) > - Storage cloning for LVM and Disk backends(Cole Robinson) > - Switching to GIT (Jim Meyering) > > Documentation: > - Typo and comment fixes (Aron Griffis) > - Fix virCapabilitiesDefaultGuestMachine documentation. (Chris Lalancette) > - ESX Scheduler documentation and cleanup (Matthias Bolte) > - Update the java bindings page (Bryan Kearney) > - Added Matthias Bolte to AUTHORS list (Daniel Veillard) > - doc: clone+build instructions (Jim Meyering) > - docs: say that the old repository is deprecated... (Jim Meyering) > - document tcp listen and raw wire option (Guido Günther) > - Fix docs and code disagreements for character devices. (Cole Robinson) > - Fix documentation of virStoragePoolUndefine return (Thomas Treutner) > - Fix gitweb link on download page. (Cole Robinson) > - update download informations after switch to git (Daniel Veillard) > - Update links to bugzilla (Garry Dolley) > - Update the links for RHEL libvirt bugzillas (Garry Dolley) > > Portability: > - Xen Inotify support needs sys/inotify.h (Maximilian Wilhelm) > - Workaround for broken GCC in Debian Etch (Maximilian Wilhelm) > - LXC driver requires sched.h and unshare() (Maximilian Wilhelm) > - Configure UML support only if sys/inotify.h present (Maximilian Wilhelm) > - Fix libcurl automatic check and ESX status (Maximilian Wilhelm) > - Enable ESX driver build on Mingw32 (Daniel P. Berrange) > - Fix build on mingw32 by disabling netcf (Daniel P. Berrange) > - Reduce glusterfs dependency to 2.0.1 (Mark McLoughlin) > - Desactivate phyp build and indicate libssh builreq (Daniel Veillard) > - Fix misc Win32 compile warnings (Daniel P. Berrange) > - Rename variable for compilation in Mingw32 (end) (Laine Stump) > - Rename variable for compilation in Mingw32 (Laine Stump) > - rpm spec cleanup and split off client only package (Daniel Veillard) > > Bug fixes: > - Add uniqueness checking for LXC define/create methods (Daniel P. Berrange) > - Fix removal of transient VMs when LXC aborts (Daniel P. Berrange) > - Don't try to activate cgroups if not present for LXC (Daniel P. Berrange) > - Refresh /etc/xen if inotify wasn't (Cole Robinson) > - Don't loose id on xen domain redefine (Cole Robinson) > - Fix memory leak in openvz driver (Daniel P. Berrange) > - Protected against potential crash scenarios (Daniel P. Berrange) > - Fix crash when attempting to shutdown inactive QEMU vm (Daniel P. Berrange) > - Fix PCIe FLR detection (Mark McLoughlin) > - Set perms on /var/lib/libvirt/images to 0711 (Mark McLoughlin) > - Fix problem writing QEMU pidfile (Daniel P. Berrange) > - Fix vcpupin on Xen problem (Henrik Persson) > - Fix RPM upgrades from F11 to F12 (Daniel P. Berrange) > - Fix deadlock in remote driver domain events (Daniel P. Berrange) > - qemu: fix monitor socket reconnection (Ryota Ozaki) > - Fix polkit/netcf disabling on older fedoras (Mark McLoughlin) > - Fix crashes in Xen capabilities code (Daniel P. Berrange) > - Always add -no-kvm and -no-kqemu, for qemu domains (Jim Paris) > - Avoid raising an internal error (Paolo Bonzini) > - Don't allow NULL paths for BlockStats and InterfaceStats (Cole Robinson) > - Don't leak vm->monitorpath on re-connect (Mark McLoughlin) > - Don't restore labels on shared/readonly disks (Daniel P. Berrange) > - Ensure spawned children have a stderr/out set to /dev/null if requested (Daniel P. Berrange) > - Ensure test:/// URIs get routed to the non-privileged libvirtd (Daniel P. Berrange) > - fix another failing "make distcheck" (qemuhelptest) (Jim Meyering) > - Fix an uninitialized variable in Unix socket open (Jun Koi) > - Fix configure flags in spec file (Daniel Veillard) > - Fix error reporting for security driver over remote protocol (Daniel P. Berrange) > - fix failing "make distcheck" (Jim Meyering) > - Fix free of unitialized data upon PCI open fail (Daniel P. Berrange) > - Fix informations about previous git server (Daniel Veillard) > - Fix memory leaks in esxDomainDumpXML (Matthias Bolte) > - Fix multiple memory leaks in virsh (Laine Stump) > - Fix PCI device hotplug/unplug with newer QEMU (Daniel P. Berrange) > - Fix problem with QEMU monitor welcome prompt confusing libvirt after a libvirtd daemon restart with active guests (Daniel P. Berrange) > - Fix python examples to use read-write conn (Dan Kenigsberg) > - Fix reconnect bug for VBox (Pritesh Kothari) > - Fix SELinux denial during hotplug (Daniel P. Berrange) > - Fix typo in check for glusterfs format pools (Daniel P. Berrange) > - Fix typo in storage cloning (Daniel P. Berrange) > - qemu: Check driver is initialized up front, to avoid segfault. (Cole Robinson) > - qemu: Try multiple times to open unix monitor socket (Cole Robinson) > - Release conn lock before reporting errors (end) (Laine Stump) > - Release conn lock before reporting interface errors (Laine Stump) > - Remove the network backend if NIC hotplug fails (Mark McLoughlin) > - Set specific flags for glusterfs fs mounts (Harshavardhana) > - storage: disk: Default to 'ext2' for new volumes. (Cole Robinson) > - storage: disk: Fix parthelper '-g' option handling. (Cole Robinson) > - storage: disk: Fix segfault creating volume without target path (Cole Robinson) > - storage: Fix deadlock when cloning across pools. (Cole Robinson) > - Update modified mac address in place in virGetInterface (Laine Stump) > > Improvements: > - Add an allocation unit when calling qemu-img (Ryota Ozaki) > - Improve diagnostics when pidfile writing fails (Daniel P. Berrange) > - Disable IPv6 on virtual networks (Daniel P. Berrange) > - Allow dnsmasq to provide DNS without DHCP (Daniel P. Berrange) > - Fix an initialization problem in previous patch (Aron Griffis) > - Remove MAX_TAP_ID and let kernel do numbering (Aron Griffis) > - Kernel command line support for UML (Ron Yorston) > - Activate the interface drivers, and cleanups (Daniel Veillard) > - Add an error code for conflicting mac addresses (Laine Stump) > - Add a test interface driver (Laine Stump) > - Add canonical machine name to capabilities output (Mark McLoughlin) > - add cd and pwd commands to virsh (Paolo Bonzini) > - Add checks for some NIC hotplug related features added in qemu-0.10.0 (Mark McLoughlin) > - Add domain autostart for LXC driver (Daniel P. Berrange) > - Add domain events support to LXC driver (Daniel P. Berrange) > - Add interface object list manipulation functions (Laine Stump) > - Add internal XML parsing/formatting flag (Mark McLoughlin) > - Add netcf XML schemas and test data (Daniel Veillard) > - Add new net filesystem glusterfs (Harshavardhana) > - Add NIC and hostnet names to domain state XML (Mark McLoughlin) > - Add no_verify query parameter to ESX URIs (Matthias Bolte) > - Add SCM_RIGHTS support to QEMU monitor code (Mark McLoughlin) > - Add support for attaching network/bridge NICs in QEMU driver (Mark McLoughlin) > - add support for netcf XML import and export (Daniel Veillard) > - Add support for network device detach (Mark McLoughlin) > - Add support for physical memory access for QEmu (Nguyen Anh Quynh) > - Add support for VBox 3 and event callbacks on vbox (Pritesh Kothari) > - Add the monitor type to the domain state XML (Mark McLoughlin) > - Add virCapsGuestMachine structure (Mark McLoughlin) > - Add virsh commands for network interface management (Laine Stump) > - Allow autostart of libvirtd to be disabled with LIBVIRT_AUTOSTART=0 (Daniel P. Berrange) > - Allow leading dots in VMX config entry names (Matthias Bolte) > - Assign names to qemu NICs and network backends (Mark McLoughlin) > - Basic qemu NIC hotplug support (Mark McLoughlin) > - build: adjust aclocal's search patch to prefer gnulib's m4 files. (Jim Meyering) > - build: automatically rerun ./bootstrap when needed (Jim Meyering) > - build: make autogen.sh use autoreconf -if (Jim Meyering) > - build: submodule machinery now works also when no tag is reachable (Jim Meyering) > - Canonicalize qemu machine types (Mark McLoughlin) > - Change code generator to give async event messages their own postfix (Daniel P. Berrange) > - Change the way client event loop watches are managed (Daniel P. Berrange) > - Decode incoming request header before invoking dispatch code (Daniel P. Berrange) > - Define an API for registering incoming message dispatch filters (Daniel P. Berrange) > - ESX driver accept VI API version 4.0 (Matthias Bolte) > - Fill in vCPU<-> pCPU current mapping, and vCPU cpuTime for QEMU (Daniel P. Berrange) > - generate ChangeLog from git logs into distribution tarball (Jim Meyering) > - Implement qemu dump capabilities (Paolo Bonzini) > - Implement schedular tunables API using cgroups (Daniel P. Berrange) > - Implement the new virinterface functions (Laine Stump) > - Make cgroups a little more efficient (Daniel P. Berrange) > - Make it easier to debug tests running programs (Daniel P. Berrange) > - Make qemuBuildHostNetStr() take tapfd as a string (Mark McLoughlin) > - Make QEMU cgroups use configurable (Daniel P. Berrange) > - Make qemuCmdFlags available in qemudDomainAttachDevice() (Mark McLoughlin) > - Move queuing of RPC replies into dispatch code (Daniel P. Berrange) > - Move vnet_hdr logic into qemudNetworkIfaceConnect() and export it (Mark McLoughlin) > - Netcf based interface driver implementation (Laine Stump) > - netcf XML validation and input and output tests (Daniel Veillard) > - Only probe qemu for machine types when binary changes (Mark McLoughlin) > - Place every QEMU guest in a private cgroup (Daniel P. Berrange) > - Probe for QEMU machine types (Mark McLoughlin) > - Probe QEMU directly for machine aliases if not found in capabilties (Mark McLoughlin) > - Public API for new virInterface functions (Laine Stump) > - python: Raise exceptions if virDomain*Stats fail. (Cole Robinson) > - Refactor cgroups to allow a group per driver to be managed directly (Daniel P. Berrange) > - Re-factor pci_add reply parsing and parse domain/bus numbers (Mark McLoughlin) > - Retain disk PCI address across libvirtd restarts (Mark McLoughlin) > - Retain PCI address from NIC attach (Mark McLoughlin) > - Run QEMU guests as an unprivileged user (Daniel P. Berrange) > - Separate code for encoding outgoing remote message headers (Daniel P. Berrange) > - Split generic RPC message dispatch code out from remote protocol API handlers (Daniel P. Berrange) > - Split out code for handling incoming method call messages (Daniel P. Berrange) > - storage: Break out actual raw cloning to separate function. (Cole Robinson) > - storage: cleanup: do away with 'createFile' (Cole Robinson) > - storage: disk: Use capacity, not allocation, when creating volume. (Cole Robinson) > - storage: Don't try sparse detection if writing to block device. (Cole Robinson) > - storage: Implement 'CreateBlockFrom' helper. (Cole Robinson) > - storage: Implement CreateVolFrom for logical and disk backend. (Cole Robinson) > - storage: Move most of the FS creation functions to common backend. (Cole Robinson) > - storage: Refactor FS backend 'create' function choosing. (Cole Robinson) > - Store the interface vlan number in the domain state (Mark McLoughlin) > - Support<video> element for QEMU guests (Daniel P. Berrange) > - Support<video> tag for defining VGA card properties (Daniel P. Berrange) > - Switch to using a unix socket for the qemu monitor (Mark McLoughlin) > - test: Generate net interface names when assigning XML. (Cole Robinson) > - test: Implement BlockStats and InterfaceStats (Cole Robinson) > - Use cgroups for block device whitelisting in QEMU guests (Daniel P. Berrange) > - Use enums for cgroup controller types / labels (Daniel P. Berrange) > - Use sendmsg() on QEMU monitor socket (Mark McLoughlin) > - Use virDomainChrTypeFromString() instead of open coding (Mark McLoughlin) > - Use virFileReadAll/virFileWriteStr for key cgroup read/write helpers (Daniel P. Berrange) > - virGetinterface matching of MAC and interface name (Laine Stump) > > Cleanups: > - Fix configure checks from previous commits (Daniel P. Berrange) > - Avoid a warning if compiling without inotify (Daniel P. Berrange) > - Remove a stray semicolon (Daniel Veillard) > - Extend the ESX URL to habdle ports and GSX (Matthias Bolte) > - Fix escaping of 8-bit high characters (Daniel P. Berrange) > - Remove ATTRIBUTE_UNUSED from flags to qemudDomainMigratePerform. (Chris Lalancette) > - Add a comment about setting errors after qemudStartVMDaemon(). (Chris Lalancette) > - Fix an erroneous debug error to KVM; it should read QEMU/KVM. (Chris Lalancette) > - Remove a stray semicolon in qemudDomainMigratePrepare2. (Chris Lalancette) > - Convert a few stray users of free() in libvirt.c to VIR_FREE(). (Chris Lalancette) > - Use virGetHostname instead of gethostname. (Chris Lalancette) > - Fix up a minor indentation issue with virDomainMigratePrepare. (Chris Lalancette) > - Fix up a silly typo in apibuild.py. (Chris Lalancette) > - Avoid warning when compiling without IFF_VNET_HDR (Maximilian Wilhelm) > - Capilize libvirt-client summary (Mark McLoughlin) > - Move ldconfig calls to libvirt-client %post/%postun (Mark McLoughlin) > - Convert NEWS to UTF-8 (Mark McLoughlin) > - Fix trailing whitespace in NEWS (Mark McLoughlin) > - No need to build require both python-devel and python (Mark McLoughlin) > - Remove executable perms from /etc/sysconfig/libvirtd (Mark McLoughlin) > - Use a %postun -p for one line scriptlet (Mark McLoughlin) > - Don't explicitly require libxml2 (Mark McLoughlin) > - Fix some unowned directories (Mark McLoughlin) > - Kill qemu BuildRequires (Mark McLoughlin) > - Enable netcf by default (Mark McLoughlin) > - Default to with_polkit (Mark McLoughlin) > - Make vbox support configurable (Mark McLoughlin) > - Build with --without-capng if capng is disabled (Mark McLoughlin) > - BuildRequires libcap-ng-devel not capng-devel (Mark McLoughlin) > - Drop curl host check when using ESX without check (Shahar Klein) > - Fix typo in xen capabilities code (Daniel P. Berrange) > - Add bare format string to printf-derivatives troubles (Laine Stump) > - Add phyp files to POTFILES, to make syntax-check happy. (Cole Robinson) > - avoid a make distcheck failure: distribute docs/schemas/interface.rng (Jim Meyering) > - avoid a make distcheck failure: distribute tests/interfaceschemadata/ (Jim Meyering) > - avoid a "make syntax-check" failure (Jim Meyering) > - build: do not emit a trailing blank line into VC'd file, NEWS (Jim Meyering) > - build: update from gnulib, for latest maint.mk (Jim Meyering) > - Clean up error handling in qemudDomainAttachNetDevice() (Mark McLoughlin) > - Cleanup qemu binary detection logic in qemudCapsInitGuest() (Mark McLoughlin) > - Commit newly generated docs, after changes from commit 2348cf. (Cole Robinson) > - Factor qemuBuildHostNetStr() out from qemuBuildCommandLine() (Mark McLoughlin) > - Factor qemuBuildNicStr() out from qemuBuildCommandLine() (Mark McLoughlin) > - Factor qemudMonitorSend() out of qemudMonitorCommandExtra() (Mark McLoughlin) > - Fix cgroup compile warnings (Daniel P. Berrange) > - Fix misc build problems due to new drivers (Daniel P. Berrange) > - Fix wierd build problems due to autopoint overwriting gnulib m4 (Daniel P. Berrange) > - Makefile.cfg: Rename to... (Jim Meyering) > - make .gnulib a submodule (Jim Meyering) > - make "make syntax-check" consistent with "git diff --check" (Jim Meyering) > - Minor qemu monitor coding style fixes (Mark McLoughlin) > - Prepare to use maint.mk from gnulib (Jim Meyering) > - Refactor incoming message handling to prepare for data stream support (Daniel P. Berrange) > - Refactor message sending to allow code reuse for data streams (Daniel P. Berrange) > - remove all .cvsignore files (Jim Meyering) > - remove all trailing blank lines (Jim Meyering) > - Remove some unused variables and cut long lines (Daniel Veillard) > - Remove trailing blank lines (Daniel Veillard) > - Rename a bunch of internal methods to clarify their meaning (Daniel P. Berrange) > - Rename 'direction' to 'type' in remote_message_header (Daniel P. Berrange) > - Report the object name on lookup error (Daniel Veillard) > - Simplify remote driver error reporting (Daniel P. Berrange) > - skip some of gnulib's new rules (Jim Meyering) > - use automake-1.11's silent-rules option, when possible (Jim Meyering) > - use gnumakefile and maintainer-makefile modules from gnulib (Jim Meyering) > > > It's also good to see a lot of new contributors in this release, > so thanks to everybody ! > > Daniel > >

2 1

[libvirt] VMware ESX 2/3: Fix memleak if esxOpen fails
by Matthias Bolte 06 Aug '09

06 Aug '09

Subject says it all. Matthias

2 1

[libvirt] VMware ESX 1/3: Add proper OOM reporting for esxDomainGetOSType
by Matthias Bolte 06 Aug '09

06 Aug '09

Subject says it all. Matthias

2 1

[libvirt] running with UID != 0: internal error about bridged network
by Harald Dunkel 06 Aug '09

06 Aug '09

Hi folks, I would like to avoid running the virtual hosts with UID=0. But if I try to create a new virtual machine without being root, then it fails with % name=Lenny_amd64 % kvmdir=/local/kvm % mkdir -p $kvmdir/$name % qemu-img create -f qcow2 $kvmdir/$name/hda.qcow2 32G Formatting '/local/kvm/Lenny_amd64/hda.qcow2', fmt=qcow2, size=33554432 kB % virt-install --connect=qemu:///session -n $name -r 1024 -vcpus=1 -f $kvmdir/$name/hda.qcow2 --network=bridge:br0 --vnc --accelerate -v -c /usr/local/source/hdunkel/debian-502-i386-netinst.iso --os-type=linux --os-variant=debianLenny 14:18:35.586: warning : Cannot set group when not running as root 14:18:35.757: error : Domain not found: no domain with matching name 'Lenny_amd64' libvir: QEMU error : Domain not found: no domain with matching name 'Lenny_amd64' The log file shows [Wed, 05 Aug 2009 14:18:35 virt-install 14229] ERROR (virt-install:786) internal error Failed to add tap interface 'vnet%d' to bridge 'br0' : Operation not permitted Traceback (most recent call last): File "/usr/bin/virt-install", line 780, in <module> main() File "/usr/bin/virt-install", line 678, in main start_time, guest.start_install) File "/usr/bin/virt-install", line 733, in do_install dom = install_func(conscb, progresscb, wait=(not wait)) File "/var/lib/python-support/python2.5/virtinst/Guest.py", line 541, in start_install return self._do_install(consolecb, meter, removeOld, wait) File "/var/lib/python-support/python2.5/virtinst/Guest.py", line 633, in _do_install self.domain = self.conn.createLinux(install_xml, 0) File "/usr/lib/python2.5/site-packages/libvirt.py", line 892, in createLinux if ret is None:raise libvirtError('virDomainCreateLinux() failed', conn=self) libvirtError: internal error Failed to add tap interface 'vnet%d' to bridge 'br0' : Operation not permitted Of course an unprivileged user is not allowed to mess around with the network interfaces, but since it says "internal error" I wonder how this is supposed to work? Is there something misconfigured? Any helpful comments would be highly appreciated Harri

2 1

[libvirt] [PATCH 0/7] QEMU network/bridge backed NIC hotplug
by Mark McLoughlin 05 Aug '09

05 Aug '09

Hey, So, here are the patches needed to complete the QEMU NIC hotplug work by supporting network and bridge backed interfaces. We support this by passing a tap file descriptor to qemu over the monitor socket using SCM_RIGHTS. The monitor commands required for this (getfd and closefd) are queued up in Anthony's queue and should be in qemu-0.11-rc1 which is due to be released very soon: http://repo.or.cz/w/qemu/aliguori-queue.git?a=commitdiff;h=4edd6c2645 http://repo.or.cz/w/qemu/aliguori-queue.git?a=commitdiff;h=f4069061e3 Note: I just remembered I'm not handling missing support for these commands very well. I'll make it handle an 'unknown command' reply gracefully by failing the attach tommorrow. Cheers, Mark.

6 41

[libvirt] Release of libvirt-0.7.0
by Daniel Veillard 05 Aug '09

05 Aug '09

A couple days later than expected, but considering the current flow of fixes, that's not a bad thing. This is a huge release, this includes more than 250 commits in a month and many new functionalities or drivers, and a lot of improvements and bug fixes: New features: - Interface implementation based on netcf (Laine Stump Daniel Veillard) - Add new net filesystem glusterfs (Harshavardhana) - Initial VMWare ESX driver (Matthias Bolte) - Add support for VBox 3 and event callbacks on vbox (Pritesh Kothari) - First version of the Power Hypervisor driver (Eduardo Otubo) - Run QEMU guests as an unprivileged user (Daniel P. Berrange) - Support cgroups in QEMU driver (Daniel P. Berrange) - QEmu hotplug NIC support (Mark McLoughlin) - Storage cloning for LVM and Disk backends(Cole Robinson) - Switching to GIT (Jim Meyering) Documentation: - Typo and comment fixes (Aron Griffis) - Fix virCapabilitiesDefaultGuestMachine documentation. (Chris Lalancette) - ESX Scheduler documentation and cleanup (Matthias Bolte) - Update the java bindings page (Bryan Kearney) - Added Matthias Bolte to AUTHORS list (Daniel Veillard) - doc: clone+build instructions (Jim Meyering) - docs: say that the old repository is deprecated... (Jim Meyering) - document tcp listen and raw wire option (Guido Günther) - Fix docs and code disagreements for character devices. (Cole Robinson) - Fix documentation of virStoragePoolUndefine return (Thomas Treutner) - Fix gitweb link on download page. (Cole Robinson) - update download informations after switch to git (Daniel Veillard) - Update links to bugzilla (Garry Dolley) - Update the links for RHEL libvirt bugzillas (Garry Dolley) Portability: - Xen Inotify support needs sys/inotify.h (Maximilian Wilhelm) - Workaround for broken GCC in Debian Etch (Maximilian Wilhelm) - LXC driver requires sched.h and unshare() (Maximilian Wilhelm) - Configure UML support only if sys/inotify.h present (Maximilian Wilhelm) - Fix libcurl automatic check and ESX status (Maximilian Wilhelm) - Enable ESX driver build on Mingw32 (Daniel P. Berrange) - Fix build on mingw32 by disabling netcf (Daniel P. Berrange) - Reduce glusterfs dependency to 2.0.1 (Mark McLoughlin) - Desactivate phyp build and indicate libssh builreq (Daniel Veillard) - Fix misc Win32 compile warnings (Daniel P. Berrange) - Rename variable for compilation in Mingw32 (end) (Laine Stump) - Rename variable for compilation in Mingw32 (Laine Stump) - rpm spec cleanup and split off client only package (Daniel Veillard) Bug fixes: - Add uniqueness checking for LXC define/create methods (Daniel P. Berrange) - Fix removal of transient VMs when LXC aborts (Daniel P. Berrange) - Don't try to activate cgroups if not present for LXC (Daniel P. Berrange) - Refresh /etc/xen if inotify wasn't (Cole Robinson) - Don't loose id on xen domain redefine (Cole Robinson) - Fix memory leak in openvz driver (Daniel P. Berrange) - Protected against potential crash scenarios (Daniel P. Berrange) - Fix crash when attempting to shutdown inactive QEMU vm (Daniel P. Berrange) - Fix PCIe FLR detection (Mark McLoughlin) - Set perms on /var/lib/libvirt/images to 0711 (Mark McLoughlin) - Fix problem writing QEMU pidfile (Daniel P. Berrange) - Fix vcpupin on Xen problem (Henrik Persson) - Fix RPM upgrades from F11 to F12 (Daniel P. Berrange) - Fix deadlock in remote driver domain events (Daniel P. Berrange) - qemu: fix monitor socket reconnection (Ryota Ozaki) - Fix polkit/netcf disabling on older fedoras (Mark McLoughlin) - Fix crashes in Xen capabilities code (Daniel P. Berrange) - Always add -no-kvm and -no-kqemu, for qemu domains (Jim Paris) - Avoid raising an internal error (Paolo Bonzini) - Don't allow NULL paths for BlockStats and InterfaceStats (Cole Robinson) - Don't leak vm->monitorpath on re-connect (Mark McLoughlin) - Don't restore labels on shared/readonly disks (Daniel P. Berrange) - Ensure spawned children have a stderr/out set to /dev/null if requested (Daniel P. Berrange) - Ensure test:/// URIs get routed to the non-privileged libvirtd (Daniel P. Berrange) - fix another failing "make distcheck" (qemuhelptest) (Jim Meyering) - Fix an uninitialized variable in Unix socket open (Jun Koi) - Fix configure flags in spec file (Daniel Veillard) - Fix error reporting for security driver over remote protocol (Daniel P. Berrange) - fix failing "make distcheck" (Jim Meyering) - Fix free of unitialized data upon PCI open fail (Daniel P. Berrange) - Fix informations about previous git server (Daniel Veillard) - Fix memory leaks in esxDomainDumpXML (Matthias Bolte) - Fix multiple memory leaks in virsh (Laine Stump) - Fix PCI device hotplug/unplug with newer QEMU (Daniel P. Berrange) - Fix problem with QEMU monitor welcome prompt confusing libvirt after a libvirtd daemon restart with active guests (Daniel P. Berrange) - Fix python examples to use read-write conn (Dan Kenigsberg) - Fix reconnect bug for VBox (Pritesh Kothari) - Fix SELinux denial during hotplug (Daniel P. Berrange) - Fix typo in check for glusterfs format pools (Daniel P. Berrange) - Fix typo in storage cloning (Daniel P. Berrange) - qemu: Check driver is initialized up front, to avoid segfault. (Cole Robinson) - qemu: Try multiple times to open unix monitor socket (Cole Robinson) - Release conn lock before reporting errors (end) (Laine Stump) - Release conn lock before reporting interface errors (Laine Stump) - Remove the network backend if NIC hotplug fails (Mark McLoughlin) - Set specific flags for glusterfs fs mounts (Harshavardhana) - storage: disk: Default to 'ext2' for new volumes. (Cole Robinson) - storage: disk: Fix parthelper '-g' option handling. (Cole Robinson) - storage: disk: Fix segfault creating volume without target path (Cole Robinson) - storage: Fix deadlock when cloning across pools. (Cole Robinson) - Update modified mac address in place in virGetInterface (Laine Stump) Improvements: - Add an allocation unit when calling qemu-img (Ryota Ozaki) - Improve diagnostics when pidfile writing fails (Daniel P. Berrange) - Disable IPv6 on virtual networks (Daniel P. Berrange) - Allow dnsmasq to provide DNS without DHCP (Daniel P. Berrange) - Fix an initialization problem in previous patch (Aron Griffis) - Remove MAX_TAP_ID and let kernel do numbering (Aron Griffis) - Kernel command line support for UML (Ron Yorston) - Activate the interface drivers, and cleanups (Daniel Veillard) - Add an error code for conflicting mac addresses (Laine Stump) - Add a test interface driver (Laine Stump) - Add canonical machine name to capabilities output (Mark McLoughlin) - add cd and pwd commands to virsh (Paolo Bonzini) - Add checks for some NIC hotplug related features added in qemu-0.10.0 (Mark McLoughlin) - Add domain autostart for LXC driver (Daniel P. Berrange) - Add domain events support to LXC driver (Daniel P. Berrange) - Add interface object list manipulation functions (Laine Stump) - Add internal XML parsing/formatting flag (Mark McLoughlin) - Add netcf XML schemas and test data (Daniel Veillard) - Add new net filesystem glusterfs (Harshavardhana) - Add NIC and hostnet names to domain state XML (Mark McLoughlin) - Add no_verify query parameter to ESX URIs (Matthias Bolte) - Add SCM_RIGHTS support to QEMU monitor code (Mark McLoughlin) - Add support for attaching network/bridge NICs in QEMU driver (Mark McLoughlin) - add support for netcf XML import and export (Daniel Veillard) - Add support for network device detach (Mark McLoughlin) - Add support for physical memory access for QEmu (Nguyen Anh Quynh) - Add support for VBox 3 and event callbacks on vbox (Pritesh Kothari) - Add the monitor type to the domain state XML (Mark McLoughlin) - Add virCapsGuestMachine structure (Mark McLoughlin) - Add virsh commands for network interface management (Laine Stump) - Allow autostart of libvirtd to be disabled with LIBVIRT_AUTOSTART=0 (Daniel P. Berrange) - Allow leading dots in VMX config entry names (Matthias Bolte) - Assign names to qemu NICs and network backends (Mark McLoughlin) - Basic qemu NIC hotplug support (Mark McLoughlin) - build: adjust aclocal's search patch to prefer gnulib's m4 files. (Jim Meyering) - build: automatically rerun ./bootstrap when needed (Jim Meyering) - build: make autogen.sh use autoreconf -if (Jim Meyering) - build: submodule machinery now works also when no tag is reachable (Jim Meyering) - Canonicalize qemu machine types (Mark McLoughlin) - Change code generator to give async event messages their own postfix (Daniel P. Berrange) - Change the way client event loop watches are managed (Daniel P. Berrange) - Decode incoming request header before invoking dispatch code (Daniel P. Berrange) - Define an API for registering incoming message dispatch filters (Daniel P. Berrange) - ESX driver accept VI API version 4.0 (Matthias Bolte) - Fill in vCPU <-> pCPU current mapping, and vCPU cpuTime for QEMU (Daniel P. Berrange) - generate ChangeLog from git logs into distribution tarball (Jim Meyering) - Implement qemu dump capabilities (Paolo Bonzini) - Implement schedular tunables API using cgroups (Daniel P. Berrange) - Implement the new virinterface functions (Laine Stump) - Make cgroups a little more efficient (Daniel P. Berrange) - Make it easier to debug tests running programs (Daniel P. Berrange) - Make qemuBuildHostNetStr() take tapfd as a string (Mark McLoughlin) - Make QEMU cgroups use configurable (Daniel P. Berrange) - Make qemuCmdFlags available in qemudDomainAttachDevice() (Mark McLoughlin) - Move queuing of RPC replies into dispatch code (Daniel P. Berrange) - Move vnet_hdr logic into qemudNetworkIfaceConnect() and export it (Mark McLoughlin) - Netcf based interface driver implementation (Laine Stump) - netcf XML validation and input and output tests (Daniel Veillard) - Only probe qemu for machine types when binary changes (Mark McLoughlin) - Place every QEMU guest in a private cgroup (Daniel P. Berrange) - Probe for QEMU machine types (Mark McLoughlin) - Probe QEMU directly for machine aliases if not found in capabilties (Mark McLoughlin) - Public API for new virInterface functions (Laine Stump) - python: Raise exceptions if virDomain*Stats fail. (Cole Robinson) - Refactor cgroups to allow a group per driver to be managed directly (Daniel P. Berrange) - Re-factor pci_add reply parsing and parse domain/bus numbers (Mark McLoughlin) - Retain disk PCI address across libvirtd restarts (Mark McLoughlin) - Retain PCI address from NIC attach (Mark McLoughlin) - Run QEMU guests as an unprivileged user (Daniel P. Berrange) - Separate code for encoding outgoing remote message headers (Daniel P. Berrange) - Split generic RPC message dispatch code out from remote protocol API handlers (Daniel P. Berrange) - Split out code for handling incoming method call messages (Daniel P. Berrange) - storage: Break out actual raw cloning to separate function. (Cole Robinson) - storage: cleanup: do away with 'createFile' (Cole Robinson) - storage: disk: Use capacity, not allocation, when creating volume. (Cole Robinson) - storage: Don't try sparse detection if writing to block device. (Cole Robinson) - storage: Implement 'CreateBlockFrom' helper. (Cole Robinson) - storage: Implement CreateVolFrom for logical and disk backend. (Cole Robinson) - storage: Move most of the FS creation functions to common backend. (Cole Robinson) - storage: Refactor FS backend 'create' function choosing. (Cole Robinson) - Store the interface vlan number in the domain state (Mark McLoughlin) - Support <video> element for QEMU guests (Daniel P. Berrange) - Support <video> tag for defining VGA card properties (Daniel P. Berrange) - Switch to using a unix socket for the qemu monitor (Mark McLoughlin) - test: Generate net interface names when assigning XML. (Cole Robinson) - test: Implement BlockStats and InterfaceStats (Cole Robinson) - Use cgroups for block device whitelisting in QEMU guests (Daniel P. Berrange) - Use enums for cgroup controller types / labels (Daniel P. Berrange) - Use sendmsg() on QEMU monitor socket (Mark McLoughlin) - Use virDomainChrTypeFromString() instead of open coding (Mark McLoughlin) - Use virFileReadAll/virFileWriteStr for key cgroup read/write helpers (Daniel P. Berrange) - virGetinterface matching of MAC and interface name (Laine Stump) Cleanups: - Fix configure checks from previous commits (Daniel P. Berrange) - Avoid a warning if compiling without inotify (Daniel P. Berrange) - Remove a stray semicolon (Daniel Veillard) - Extend the ESX URL to habdle ports and GSX (Matthias Bolte) - Fix escaping of 8-bit high characters (Daniel P. Berrange) - Remove ATTRIBUTE_UNUSED from flags to qemudDomainMigratePerform. (Chris Lalancette) - Add a comment about setting errors after qemudStartVMDaemon(). (Chris Lalancette) - Fix an erroneous debug error to KVM; it should read QEMU/KVM. (Chris Lalancette) - Remove a stray semicolon in qemudDomainMigratePrepare2. (Chris Lalancette) - Convert a few stray users of free() in libvirt.c to VIR_FREE(). (Chris Lalancette) - Use virGetHostname instead of gethostname. (Chris Lalancette) - Fix up a minor indentation issue with virDomainMigratePrepare. (Chris Lalancette) - Fix up a silly typo in apibuild.py. (Chris Lalancette) - Avoid warning when compiling without IFF_VNET_HDR (Maximilian Wilhelm) - Capilize libvirt-client summary (Mark McLoughlin) - Move ldconfig calls to libvirt-client %post/%postun (Mark McLoughlin) - Convert NEWS to UTF-8 (Mark McLoughlin) - Fix trailing whitespace in NEWS (Mark McLoughlin) - No need to build require both python-devel and python (Mark McLoughlin) - Remove executable perms from /etc/sysconfig/libvirtd (Mark McLoughlin) - Use a %postun -p for one line scriptlet (Mark McLoughlin) - Don't explicitly require libxml2 (Mark McLoughlin) - Fix some unowned directories (Mark McLoughlin) - Kill qemu BuildRequires (Mark McLoughlin) - Enable netcf by default (Mark McLoughlin) - Default to with_polkit (Mark McLoughlin) - Make vbox support configurable (Mark McLoughlin) - Build with --without-capng if capng is disabled (Mark McLoughlin) - BuildRequires libcap-ng-devel not capng-devel (Mark McLoughlin) - Drop curl host check when using ESX without check (Shahar Klein) - Fix typo in xen capabilities code (Daniel P. Berrange) - Add bare format string to printf-derivatives troubles (Laine Stump) - Add phyp files to POTFILES, to make syntax-check happy. (Cole Robinson) - avoid a make distcheck failure: distribute docs/schemas/interface.rng (Jim Meyering) - avoid a make distcheck failure: distribute tests/interfaceschemadata/ (Jim Meyering) - avoid a "make syntax-check" failure (Jim Meyering) - build: do not emit a trailing blank line into VC'd file, NEWS (Jim Meyering) - build: update from gnulib, for latest maint.mk (Jim Meyering) - Clean up error handling in qemudDomainAttachNetDevice() (Mark McLoughlin) - Cleanup qemu binary detection logic in qemudCapsInitGuest() (Mark McLoughlin) - Commit newly generated docs, after changes from commit 2348cf. (Cole Robinson) - Factor qemuBuildHostNetStr() out from qemuBuildCommandLine() (Mark McLoughlin) - Factor qemuBuildNicStr() out from qemuBuildCommandLine() (Mark McLoughlin) - Factor qemudMonitorSend() out of qemudMonitorCommandExtra() (Mark McLoughlin) - Fix cgroup compile warnings (Daniel P. Berrange) - Fix misc build problems due to new drivers (Daniel P. Berrange) - Fix wierd build problems due to autopoint overwriting gnulib m4 (Daniel P. Berrange) - Makefile.cfg: Rename to... (Jim Meyering) - make .gnulib a submodule (Jim Meyering) - make "make syntax-check" consistent with "git diff --check" (Jim Meyering) - Minor qemu monitor coding style fixes (Mark McLoughlin) - Prepare to use maint.mk from gnulib (Jim Meyering) - Refactor incoming message handling to prepare for data stream support (Daniel P. Berrange) - Refactor message sending to allow code reuse for data streams (Daniel P. Berrange) - remove all .cvsignore files (Jim Meyering) - remove all trailing blank lines (Jim Meyering) - Remove some unused variables and cut long lines (Daniel Veillard) - Remove trailing blank lines (Daniel Veillard) - Rename a bunch of internal methods to clarify their meaning (Daniel P. Berrange) - Rename 'direction' to 'type' in remote_message_header (Daniel P. Berrange) - Report the object name on lookup error (Daniel Veillard) - Simplify remote driver error reporting (Daniel P. Berrange) - skip some of gnulib's new rules (Jim Meyering) - use automake-1.11's silent-rules option, when possible (Jim Meyering) - use gnumakefile and maintainer-makefile modules from gnulib (Jim Meyering) It's also good to see a lot of new contributors in this release, so thanks to everybody ! Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

1 0

[libvirt] [PATCH] address incompatible change of qemu-img
by Ryota Ozaki 05 Aug '09

05 Aug '09

Hi, Latest qemu-img (maybe from 0.11.0-rc0 or later?) changes default unit of size as bytes, it was kilo-bytes so far. (see http://www.mail-archive.com/kvm@vger.kernel.org/msg19183.html) This patch addresses this incompatibility by adding postfix 'K'. This fix is able to keep backward compatibility for old qemu as well, and also even if qemu-img falls back old style or any, this explicit specifying should work well. Thanks, ozaki-r >From 7d3dd9e592deab5a1d793532cf230d8367372280 Mon Sep 17 00:00:00 2001 From: Ryota Ozaki <ozaki.ryota(a)gmail.com> Date: Tue, 4 Aug 2009 11:11:31 +0900 Subject: [PATCH] address incompatible change of qemu-img --- src/storage_backend.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/storage_backend.c b/src/storage_backend.c index 07a2e48..8824218 100644 --- a/src/storage_backend.c +++ b/src/storage_backend.c @@ -446,7 +446,7 @@ virStorageBackendCreateQemuImg(virConnectPtr conn, /* Size in KB */ - snprintf(size, sizeof(size), "%llu", vol->capacity/1024); + snprintf(size, sizeof(size), "%lluK", vol->capacity/1024); if (virRun(conn, imgargv, NULL) < 0) { VIR_FREE(imgargv[0]); -- 1.6.0.6

2 1

[libvirt] [PATCH] Support the usage of socat instead of netcat
by Jonas Eriksson 05 Aug '09

05 Aug '09

Some netcat implementations does not support unix domain sockets (-U). Distros that ship these implementations often also ships socat, that do support this. This is a reworked patch based on the openSUSE libvirt package. The original patch may be found in this source rpm: http://download.opensuse.org/source/distribution/11.1/repo/oss/suse/src/lib… * configure.in: Check for --with-socat and set USE_SOCAT in config.h * src/remote_internal.c: Honour USE_SOCAT by selecting between netcat/nc and socat at compile time. --- configure.in | 7 +++++++ src/remote_internal.c | 13 +++++++++++++ 2 files changed, 20 insertions(+), 0 deletions(-) diff --git a/configure.in b/configure.in index b905b23..732ef68 100644 --- a/configure.in +++ b/configure.in @@ -153,6 +153,13 @@ if test -n "$MODPROBE"; then [Location or name of the modprobe program]) fi +dnl Check if socat should be used instead of netcat +AC_ARG_WITH([socat], +[ --with-socat use socat instead of netcat (off)]) +if test "$with_socat" = "yes" ; then + AC_DEFINE_UNQUOTED([USE_SOCAT], 1, [whether socat should be used instead of netcat]) +fi + dnl Specific dir for HTML output ? AC_ARG_WITH([html-dir], [AC_HELP_STRING([--with-html-dir=path], [path to base html directory, default $datadir/doc/html])], diff --git a/src/remote_internal.c b/src/remote_internal.c index a58b768..7678317 100644 --- a/src/remote_internal.c +++ b/src/remote_internal.c @@ -725,6 +725,18 @@ doRemoteOpen (virConnectPtr conn, cmd_argv[j++] = strdup ("-e"); cmd_argv[j++] = strdup ("none"); } +#ifdef USE_SOCAT + cmd_argv[j++] = strdup ("socat"); + cmd_argv[j++] = strdup ("-"); + + char *socat_addr = 0; + if ((asprintf (&socat_addr, "GOPEN:%s", + sockname ? sockname : LIBVIRTD_PRIV_UNIX_SOCKET)) < 0) { + error (conn, VIR_ERR_SYSTEM_ERROR, strerror (ENOMEM)); + goto failed; + } + cmd_argv[j++] = socat_addr; +#else cmd_argv[j++] = strdup (priv->hostname); cmd_argv[j++] = strdup (netcat ? netcat : "nc"); cmd_argv[j++] = strdup ("-U"); @@ -732,6 +744,7 @@ doRemoteOpen (virConnectPtr conn, (flags & VIR_CONNECT_RO ? LIBVIRTD_PRIV_UNIX_SOCKET_RO : LIBVIRTD_PRIV_UNIX_SOCKET)); +#endif cmd_argv[j++] = 0; assert (j == nr_args); for (j = 0; j < (nr_args-1); j++) -- 1.6.2

1 1

[libvirt] PATCH: Add name uniqueness checking to LXC driver
by Daniel P. Berrange 05 Aug '09

05 Aug '09

The LXC driver define/create methods didn't yet have name uniqueness checking enabled Daniel commit 4b23a44f34392b5c6d1c5aa483cbe81b010e47c4 Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Fri Jul 31 15:25:03 2009 +0100 Add uniqness checking for LXC define/create methods * src/lxc_driver.c: Check for name & UUID uniqueness when defining or creating domains diff --git a/src/lxc_driver.c b/src/lxc_driver.c index d62c2d7..a9c4f79 100644 --- a/src/lxc_driver.c +++ b/src/lxc_driver.c @@ -311,6 +311,35 @@ static virDomainPtr lxcDomainDefine(virConnectPtr conn, const char *xml) VIR_DOMAIN_XML_INACTIVE))) goto cleanup; + /* See if a VM with matching UUID already exists */ + vm = virDomainFindByUUID(&driver->domains, def->uuid); + if (vm) { + /* UUID matches, but if names don't match, refuse it */ + if (STRNEQ(vm->def->name, def->name)) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(vm->def->uuid, uuidstr); + lxcError(conn, NULL, VIR_ERR_OPERATION_FAILED, + _("domain '%s' is already defined with uuid %s"), + vm->def->name, uuidstr); + goto cleanup; + } + + /* UUID & name match */ + virDomainObjUnlock(vm); + newVM = 0; + } else { + /* UUID does not match, but if a name matches, refuse it */ + vm = virDomainFindByName(&driver->domains, def->name); + if (vm) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(vm->def->uuid, uuidstr); + lxcError(conn, NULL, VIR_ERR_OPERATION_FAILED, + _("domain '%s' is already defined with uuid %s"), + def->name, uuidstr); + goto cleanup; + } + } + if ((def->nets != NULL) && !(driver->have_netns)) { lxcError(conn, NULL, VIR_ERR_NO_SUPPORT, "%s", _("System lacks NETNS support")); @@ -1082,6 +1111,39 @@ lxcDomainCreateAndStart(virConnectPtr conn, VIR_DOMAIN_XML_INACTIVE))) goto cleanup; + /* See if a VM with matching UUID already exists */ + vm = virDomainFindByUUID(&driver->domains, def->uuid); + if (vm) { + /* UUID matches, but if names don't match, refuse it */ + if (STRNEQ(vm->def->name, def->name)) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(vm->def->uuid, uuidstr); + lxcError(conn, NULL, VIR_ERR_OPERATION_FAILED, + _("domain '%s' is already defined with uuid %s"), + vm->def->name, uuidstr); + goto cleanup; + } + + /* UUID & name match, but if VM is already active, refuse it */ + if (virDomainIsActive(vm)) { + lxcError(conn, NULL, VIR_ERR_OPERATION_FAILED, + _("domain is already active as '%s'"), vm->def->name); + goto cleanup; + } + virDomainObjUnlock(vm); + } else { + /* UUID does not match, but if a name matches, refuse it */ + vm = virDomainFindByName(&driver->domains, def->name); + if (vm) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(vm->def->uuid, uuidstr); + lxcError(conn, NULL, VIR_ERR_OPERATION_FAILED, + _("domain '%s' is already defined with uuid %s"), + def->name, uuidstr); + goto cleanup; + } + } + if ((def->nets != NULL) && !(driver->have_netns)) { lxcError(conn, NULL, VIR_ERR_NO_SUPPORT, "%s", _("System lacks NETNS support")); -- |: 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 :|

2 1

[libvirt] PATCH: Fix cleanup of transient VMs
by Daniel P. Berrange 05 Aug '09

05 Aug '09

If a transient VM quit unexpectedly it would never be cleaned up, resulting in a guest that's impossible to remove Daniel commit 7781032c3d870f5c9dacfc02baecc151689d9c57 Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Fri Jul 31 14:38:46 2009 +0100 Fix removal of transient VMs when LXC aborts * src/lxc_driver.c: Remove transient VM after monitor triggered shutdown diff --git a/src/lxc_driver.c b/src/lxc_driver.c index f37fc5d..d62c2d7 100644 --- a/src/lxc_driver.c +++ b/src/lxc_driver.c @@ -774,6 +774,10 @@ static void lxcMonitorEvent(int watch, VIR_DOMAIN_EVENT_STOPPED, VIR_DOMAIN_EVENT_STOPPED_SHUTDOWN); } + if (!vm->persistent) { + virDomainRemoveInactive(&driver->domains, vm); + vm = NULL; + } cleanup: if (vm) -- |: 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 :|

2 1

[libvirt] PATCH: Fix startup of LXC vms without cgroups
by Daniel P. Berrange 05 Aug '09

05 Aug '09

If cgroups was not mounted at all, LXC would fail to start any VM, since we missed an error code check Daniel commit 8713c7e8111e73879f8380b920995ff3d1f3f9e9 Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Fri Jul 31 14:37:25 2009 +0100 Don't try to activate cgroups if not present for LXC * src/lxc_controller.c: Don't throw error in LXC startup if the cgroups driver mount isn't available. Improve error logging for resource setup diff --git a/src/lxc_controller.c b/src/lxc_controller.c index 9ad48a7..8d11238 100644 --- a/src/lxc_controller.c +++ b/src/lxc_controller.c @@ -84,25 +84,38 @@ static int lxcSetContainerResources(virDomainDefPtr def) rc = virCgroupForDriver("lxc", &driver, 1, 0); if (rc != 0) { - lxcError(NULL, NULL, VIR_ERR_INTERNAL_ERROR, "%s", - _("Unable to get cgroup for driver")); + /* Skip all if no driver cgroup is configured */ + if (rc == -ENXIO || rc == -ENOENT) + return 0; + + virReportSystemError(NULL, -rc, "%s", + _("Unable to get cgroup for driver")); return rc; } rc = virCgroupForDomain(driver, def->name, &cgroup, 1); if (rc != 0) { - lxcError(NULL, NULL, VIR_ERR_INTERNAL_ERROR, - _("Unable to create cgroup for domain %s"), def->name); + virReportSystemError(NULL, -rc, + _("Unable to create cgroup for domain %s"), + def->name); goto cleanup; } rc = virCgroupSetMemory(cgroup, def->maxmem); - if (rc != 0) - goto out; + if (rc != 0) { + virReportSystemError(NULL, -rc, + _("Unable to set memory limit for domain %s"), + def->name); + goto cleanup; + } rc = virCgroupDenyAllDevices(cgroup); - if (rc != 0) - goto out; + if (rc != 0) { + virReportSystemError(NULL, -rc, + _("Unable to deny devices for domain %s"), + def->name); + goto cleanup; + } for (i = 0; devices[i].type != 0; i++) { struct cgroup_device_policy *dev = &devices[i]; @@ -110,19 +123,27 @@ static int lxcSetContainerResources(virDomainDefPtr def) dev->type, dev->major, dev->minor); - if (rc != 0) - goto out; + if (rc != 0) { + virReportSystemError(NULL, -rc, + _("Unable to allow device %c:%d:%d for domain %s"), + dev->type, dev->major, dev->minor, def->name); + goto cleanup; + } } rc = virCgroupAllowDeviceMajor(cgroup, 'c', LXC_DEV_MAJ_PTY); - if (rc != 0) - goto out; + if (rc != 0) { + virReportSystemError(NULL, -rc, + _("Unable to allow PYT devices for domain %s"), + def->name); + goto cleanup; + } rc = virCgroupAddTask(cgroup, getpid()); -out: if (rc != 0) { - virReportSystemError(NULL, -rc, "%s", - _("Failed to set lxc resources")); + virReportSystemError(NULL, -rc, + _("Unable to add task %d to cgroup for domain %s"), + getpid(), def->name); } cleanup: -- |: 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 :|

2 1

[libvirt] [PATCH] Refresh /etc/xen if inotify wasn't initialized
by Cole Robinson 05 Aug '09

05 Aug '09

If libvirt was built against inotify, but an API user isn't using domain events (virsh, virt-manager), the xen xm driver doesn't poll /etc/xen for new config files. This means that domains created or deleted on other libvirt connections aren't reflected in the original connection (but they will show up for all newly opened connections). This causes newly created VMs to disappear from virt-manager < 0.8.0 once they are shutdown, since VMs were installed on a separate connection. The attached patch refreshes /etc/xen in all cases, unless inotify was successfully initialized (user is using domain events). Thanks, Cole

3 3

[libvirt] [PATCH] Allow libvirtd to RPC to external libvirtd.
by Chris Lalancette 05 Aug '09

05 Aug '09

Allow the daemon itself to make RPCs to an external libvirtd, but only if the URI is fully specified. While this isn't used at the moment, it will be for the tunnelled migration support in the future. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/remote_internal.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/remote_internal.c b/src/remote_internal.c index ad0cbf0..ff4fb6c 100644 --- a/src/remote_internal.c +++ b/src/remote_internal.c @@ -997,7 +997,7 @@ remoteOpen (virConnectPtr conn, int ret, rflags = 0; const char *autostart = getenv("LIBVIRT_AUTOSTART"); - if (inside_daemon) + if (inside_daemon && (!conn->uri || (conn->uri && !conn->uri->server))) return VIR_DRV_OPEN_DECLINED; if (!(priv = remoteAllocPrivateData(conn))) -- 1.6.0.6

3 3

[libvirt] [PATCH] rng comment fixes
by Aron Griffis 05 Aug '09

05 Aug '09

--- docs/schemas/capability.rng | 3 +-- docs/schemas/nodedev.rng | 4 ---- docs/schemas/storagepool.rng | 3 +-- docs/schemas/storagevol.rng | 3 +-- 4 files changed, 3 insertions(+), 10 deletions(-) diff --git a/docs/schemas/capability.rng b/docs/schemas/capability.rng index 775bbdb..1e3c5f3 100644 --- a/docs/schemas/capability.rng +++ b/docs/schemas/capability.rng @@ -1,7 +1,6 @@ - + <grammar xmlns="http://relaxng.org/ns/structure/1.0" datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes"> -  <start> <ref name='capabilities'/> </start> diff --git a/docs/schemas/nodedev.rng b/docs/schemas/nodedev.rng index 78bd6a8..7060274 100644 --- a/docs/schemas/nodedev.rng +++ b/docs/schemas/nodedev.rng @@ -1,14 +1,10 @@  <grammar xmlns="http://relaxng.org/ns/structure/1.0" datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes"> -  <start> <ref name='device'/> </start> -  <define name='device'> <element name="device">  + <grammar xmlns="http://relaxng.org/ns/structure/1.0" datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes"> -  <start> <ref name='pool'/> </start> diff --git a/docs/schemas/storagevol.rng b/docs/schemas/storagevol.rng index 7dc7876..5b0b038 100644 --- a/docs/schemas/storagevol.rng +++ b/docs/schemas/storagevol.rng @@ -1,7 +1,6 @@ - + <grammar xmlns="http://relaxng.org/ns/structure/1.0" datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes"> -  <start> <ref name='vol'/> </start>

2 1

[libvirt] openvzGetType() leaks memory
by Matthias Bolte 05 Aug '09

05 Aug '09

All other drivers return a static string from their getType() function (conforming with the API documentation), but the OpenVZ driver does strdup() a string, generating a leak: static const char *openvzGetType(virConnectPtr conn ATTRIBUTE_UNUSED) { return strdup("OpenVZ"); } Matthias

2 1

[libvirt] [PATCH] Don't clobber ID when redefining dom in xm driver
by Cole Robinson 05 Aug '09

05 Aug '09

If redefining a domain in the xen xm driver, we unconditionally set the domain ID to -1. This can make a running domain appear inactive, and really confuse things. Just drop the offending code, since virGetDomain already sets the ID to -1 if required. Fix pointed out by Dan Berrange, more info can be found at https://bugzilla.redhat.com/show_bug.cgi?id=508278 Thanks, Cole

3 2

[libvirt] [PATCH] Fix latent buffer overflow in qemudOpenMonitorUnix.
by Chris Lalancette 05 Aug '09

05 Aug '09

Fix a possible latent bug in qemudOpenMonitorUnix(). If the pathname to the monitor is very long (i.e. >= UNIX_MAX_PATH), then strncpy will *not* place a final \0 on the string (see "man strncpy"). NULL terminate the buffer to ensure we don't run off the end. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu_driver.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 9fcc07a..4f173b7 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -910,6 +910,7 @@ qemudOpenMonitorUnix(virConnectPtr conn, memset(&addr, 0, sizeof(addr)); addr.sun_family = AF_UNIX; strncpy(addr.sun_path, monitor, sizeof(addr.sun_path)); + NUL_TERMINATE(addr.sun_path); do { ret = connect(monfd, (struct sockaddr *) &addr, sizeof(addr)); -- 1.6.0.6

4 5

[libvirt] [PATCH] Remove a stray semicolon in qemudDomainMigratePrepare2.
by Chris Lalancette 05 Aug '09

05 Aug '09

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 83cbcf3..5f4b28b 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -6274,7 +6274,7 @@ qemudDomainMigratePrepare2 (virConnectPtr dconn, char migrateFrom [64]; const char *p; virDomainEventPtr event = NULL; - int ret = -1;; + int ret = -1; int internalret; *uri_out = NULL; -- 1.6.0.6

3 3

[libvirt] [PATCH] Workaround for broken GCC in Debian Etch
by Maximilian Wilhelm 05 Aug '09

05 Aug '09

Hi again! I missed one patch for working around a GCC bug in Debian Etch regarding limit definitions. Sorry for the unstripped git-format-patch outputs, I just realised this after pressing send :( (This seems to be not my day :)) Ciao Max -- Gib Dein Bestes. Dann übertriff Dich selbst!

3 4

[libvirt] [PATCH] Add some automatic dependency checks for configure to enable/disable {ESX, UML, LXC, Xen-Inotify}
by Maximilian Wilhelm 05 Aug '09

05 Aug '09

Hi! While building libVirt on an older Debian Etch machine, I stumbled across some automatic enabled drivers which stopped configure and make very badly due to missing header files, old libraries and stuff. So I decided to improve the automagic tests so others won't face the same problem again :) Find four patches attached. Ciao Max -- Fortschritt bedeutet, einen Schritt so zu machen, daß man den nächsten auch noch machen kann.

3 2

Re: [libvirt] VMware support and Server 2.0.1
by Matthias Bolte 05 Aug '09

05 Aug '09

2009/7/30 Timo Makinen <tmakinen(a)ee.oulu.fi>: > On Thu, Jul 30, 2009 at 03:22:59AM +0200, Matthias Bolte wrote: >> 2009/7/29 Timo Makinen <tmakinen(a)ee.oulu.fi>: >> > On Wed, Jul 29, 2009 at 06:12:59PM +0200, Matthias Bolte wrote: >> >> 2009/7/27 Timo Makinen <tmakinen(a)ee.oulu.fi>: >> >> Could you do some basic testing beside listing and dumpxml, as I have >> >> currently no VMware GSX installation at hand? Just test if the >> >> following commands work as expected in virsh: >> >> >> >> - start >> >> - shutdown >> >> - reboot >> >> - suspend >> >> - resume >> >> - nodeinfo >> >> - dominfo >> >> >> >> Regards, >> >> Matthias >> > >> > All these worked out of the box. >> >> Nice! >> >> Could you test the attached patch? It adds handling for the port in >> the URI, extends the version checking for GSX 2.0 and fixes the bug in >> VMX parsing. >> >> Matthias > > Works like a charm now. :) > > - Timo > Fine! The attached second GSX patch adds a gsx:// scheme to the ESX driver, so you can connect to a GSX host using virsh -c gsx://host and the driver will select the port automatically dependent on the scheme and transport. Here gsx scheme and default transport https result in port 8333. Matthias

3 2

[libvirt] [PATCH] spelling fix
by Aron Griffis 04 Aug '09

04 Aug '09

--- src/qemu_conf.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/qemu_conf.c b/src/qemu_conf.c index 4fe4e39..6b0b404 100644 --- a/src/qemu_conf.c +++ b/src/qemu_conf.c @@ -1970,7 +1970,7 @@ int qemudBuildCommandLine(virConnectPtr conn, if (def->nvideos) { if (def->nvideos > 1) { qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, - "%s", _("only one video card is currentely supported")); + "%s", _("only one video card is currently supported")); goto error; }

1 0

[libvirt] PATCH: Fix escaping for 8-bit high characters
by Daniel P. Berrange 04 Aug '09

04 Aug '09

The virBufferEscapeString method has a broken test *cur > 0x20 which causes all 8-bit high characters to be lost as 'cur' is a signed char. This obviously dooms anyone using UTF-8 characters outside the boring old ASCII-7 range, like this guy https://bugzilla.redhat.com/show_bug.cgi?id=479517 Daniel diff --git a/src/buf.c b/src/buf.c index 259175d..c802aa2 100644 --- a/src/buf.c +++ b/src/buf.c @@ -304,7 +304,7 @@ virBufferEscapeString(const virBufferPtr buf, const char *format, const char *st *out++ = 'o'; *out++ = 's'; *out++ = ';'; - } else if ((*cur >= 0x20) || (*cur == '\n') || (*cur == '\t') || + } else if (((unsigned char)*cur >= 0x20) || (*cur == '\n') || (*cur == '\t') || (*cur == '\r')) { /* * default case, just copy ! -- |: 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 :|

2 1

[libvirt] [PATCH] Remove ATTRIBUTE_UNUSED from flags to qemudDomainMigratePerform.
by Chris Lalancette 02 Aug '09

02 Aug '09

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index eadb53f..9fcc07a 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -6424,7 +6424,7 @@ qemudDomainMigratePerform (virDomainPtr dom, const char *cookie ATTRIBUTE_UNUSED, int cookielen ATTRIBUTE_UNUSED, const char *uri, - unsigned long flags ATTRIBUTE_UNUSED, + unsigned long flags, const char *dname ATTRIBUTE_UNUSED, unsigned long resource) { -- 1.6.0.6

2 2

[libvirt] [PATCH] Add a comment about setting errors after qemudStartVMDaemon().
by Chris Lalancette 02 Aug '09

02 Aug '09

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu_driver.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index fcfbfbc..eadb53f 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -6390,6 +6390,9 @@ qemudDomainMigratePrepare2 (virConnectPtr dconn, */ snprintf (migrateFrom, sizeof (migrateFrom), "tcp:0.0.0.0:%d", this_port); if (qemudStartVMDaemon (dconn, driver, vm, migrateFrom, -1) < 0) { + /* Note that we don't set an error here because qemudStartVMDaemon + * should have already done that. + */ if (!vm->persistent) { virDomainRemoveInactive(&driver->domains, vm); vm = NULL; -- 1.6.0.6

2 2

[libvirt] [PATCH] Fix an erroneous debug error to KVM; it should read QEMU/KVM.
by Chris Lalancette 02 Aug '09

02 Aug '09

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 5f4b28b..fcfbfbc 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -6321,7 +6321,7 @@ qemudDomainMigratePrepare2 (virConnectPtr dconn, */ if (!STRPREFIX (uri_in, "tcp:")) { qemudReportError (dconn, NULL, NULL, VIR_ERR_INVALID_ARG, - "%s", _("only tcp URIs are supported for KVM migrations")); + "%s", _("only tcp URIs are supported for KVM/QEMU migrations")); goto cleanup; } -- 1.6.0.6

2 2

[libvirt] [PATCH] Convert a few stray users of free() in libvirt.c to VIR_FREE().
by Chris Lalancette 02 Aug '09

02 Aug '09

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/libvirt.c | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index 167359f..9fd864d 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -3020,7 +3020,7 @@ virDomainMigrate (virDomainPtr domain, ret = dconn->driver->domainMigratePrepare2 (dconn, &cookie, &cookielen, uri, &uri_out, flags, dname, bandwidth, dom_xml); - free (dom_xml); + VIR_FREE (dom_xml); if (ret == -1) goto done; if (uri == NULL && uri_out == NULL) { virLibConnError (conn, VIR_ERR_INTERNAL_ERROR, @@ -3062,8 +3062,8 @@ virDomainMigrate (virDomainPtr domain, } done: - free (uri_out); - free (cookie); + VIR_FREE (uri_out); + VIR_FREE (cookie); return ddomain; error: -- 1.6.0.6

2 2

[libvirt] [PATCH] Use virGetHostname instead of gethostname.
by Chris Lalancette 02 Aug '09

02 Aug '09

Fix up qemudDomainMigratePrepare2 to use virGetHostname instead of gethostname. Besides the fact that virGetHostname is far more clever, there was a latent bug in the handling that could cause a buffer overflow on a very long hostname. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu_driver.c | 9 ++++++--- 1 files changed, 6 insertions(+), 3 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index cee2164..83cbcf3 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -6270,11 +6270,12 @@ qemudDomainMigratePrepare2 (virConnectPtr dconn, virDomainDefPtr def = NULL; virDomainObjPtr vm = NULL; int this_port; - char hostname [HOST_NAME_MAX+1]; + char *hostname; char migrateFrom [64]; const char *p; virDomainEventPtr event = NULL; int ret = -1;; + int internalret; *uri_out = NULL; @@ -6300,14 +6301,16 @@ qemudDomainMigratePrepare2 (virConnectPtr dconn, if (port == QEMUD_MIGRATION_NUM_PORTS) port = 0; /* Get hostname */ - if (gethostname (hostname, HOST_NAME_MAX+1) == -1) { + if ((hostname = virGetHostname()) == NULL) { virReportSystemError (dconn, errno, "%s", _("failed to determine host name")); goto cleanup; } /* Caller frees */ - if (virAsprintf(uri_out, "tcp:%s:%d", hostname, this_port) < 0) { + internalret = virAsprintf(uri_out, "tcp:%s:%d", hostname, this_port); + VIR_FREE(hostname); + if (internalret < 0) { virReportOOMError (dconn); goto cleanup; } -- 1.6.0.6

2 2

[libvirt] [PATCH] Fix up a minor indentation issue with virDomainMigratePrepare.
by Chris Lalancette 02 Aug '09

02 Aug '09

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/libvirt.c | 14 +++++++------- 1 files changed, 7 insertions(+), 7 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index 7ecf1ec..167359f 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -3078,13 +3078,13 @@ error: */ int virDomainMigratePrepare (virConnectPtr dconn, - char **cookie, - int *cookielen, - const char *uri_in, - char **uri_out, - unsigned long flags, - const char *dname, - unsigned long bandwidth) + char **cookie, + int *cookielen, + const char *uri_in, + char **uri_out, + unsigned long flags, + const char *dname, + unsigned long bandwidth) { VIR_DEBUG("dconn=%p, cookie=%p, cookielen=%p, uri_in=%s, uri_out=%p, " "flags=%lu, dname=%s, bandwidth=%lu", dconn, cookie, cookielen, -- 1.6.0.6

2 2

[libvirt] [PATCH] Fix virCapabilitiesDefaultGuestMachine documentation.
by Chris Lalancette 02 Aug '09

02 Aug '09

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/capabilities.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/capabilities.c b/src/capabilities.c index d186961..c6766b6 100644 --- a/src/capabilities.c +++ b/src/capabilities.c @@ -567,7 +567,7 @@ virCapabilitiesDefaultGuestMachine(virCapsPtr caps, } /** - * virCapabilitiesDefaultGuestMachine: + * virCapabilitiesDefaultGuestEmulator: * @caps: capabilities to query * @ostype: OS type to search for ('xen', 'hvm') * @arch: architecture to search for -- 1.6.0.6

2 2

[libvirt] [PATCH] Fix up a silly typo in apibuild.py.
by Chris Lalancette 02 Aug '09

02 Aug '09

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- docs/apibuild.py | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/docs/apibuild.py b/docs/apibuild.py index 6fec049..84bc1ac 100755 --- a/docs/apibuild.py +++ b/docs/apibuild.py @@ -39,7 +39,7 @@ ignored_functions = { "virDomainMigratePrepare": "private function for migration", "virDomainMigratePrepare2": "private function for migration", "virDrvSupportsFeature": "private function for remote access", - "DllMain": "specific function fo Win32", + "DllMain": "specific function for Win32", } def escape(raw): -- 1.6.0.6

2 2

[libvirt] PATCH: Fix crash attempting to shutdown inactive QEMU vm
by Daniel P. Berrange 31 Jul '09

31 Jul '09

If the virDomainShutdown() op was run on an active QEMu vm, it would crash, since the def->monitor_chr was NULL. Daniel commit 49ec121fd806aaa07bbcba9668a8f9dd43dda9c4 Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Fri Jul 31 15:50:46 2009 +0100 Fix crash when attempting to shutdown inactive QEMU vm * src/qemu_driver.c: Add check that QEMU is active before attemting to shutdown. Fix error code for check in destroy method diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 26897d3..cebbafb 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -2940,6 +2940,12 @@ static int qemudDomainShutdown(virDomainPtr dom) { goto cleanup; } + if (!virDomainIsActive(vm)) { + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); + goto cleanup; + } + if (qemudMonitorCommand(vm, "system_powerdown", &info) < 0) { qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, "%s", _("shutdown operation failed")); @@ -2971,7 +2977,7 @@ static int qemudDomainDestroy(virDomainPtr dom) { goto cleanup; } if (!virDomainIsActive(vm)) { - qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, + qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID, "%s", _("domain is not running")); goto cleanup; } -- |: 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 :|

2 1

[libvirt] [PATCH] Split virDomainMigrate into functions.
by Chris Lalancette 31 Jul '09

31 Jul '09

Re-factor virDomainMigrate to split out the version 1 and version 2 protocols into their own functions. In reality, the two versions share very little in common, so forcing them together in the same function was just confusing. This will also make adding tunnelled migration easier. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/libvirt.c | 258 ++++++++++++++++++++++++++++++++++----------------------- 1 files changed, 155 insertions(+), 103 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index 9fd864d..dede2b9 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -2874,6 +2874,146 @@ error: } +static virDomainPtr +migrate_version_1 (virDomainPtr domain, + virConnectPtr dconn, + unsigned long flags, + const char *dname, + const char *uri, + unsigned long bandwidth) +{ + virDomainPtr ddomain = NULL; + char *uri_out = NULL; + char *cookie = NULL; + int cookielen = 0; + + /* Prepare the migration. + * + * The destination host may return a cookie, or leave cookie as + * NULL. + * + * The destination host MUST set uri_out if uri_in is NULL. + * + * If uri_in is non-NULL, then the destination host may modify + * the URI by setting uri_out. If it does not wish to modify + * the URI, it should leave uri_out as NULL. + */ + if (dconn->driver->domainMigratePrepare + (dconn, &cookie, &cookielen, uri, &uri_out, flags, dname, + bandwidth) == -1) + goto done; + + if (uri == NULL && uri_out == NULL) { + virLibConnError (domain->conn, VIR_ERR_INTERNAL_ERROR, + _("domainMigratePrepare did not set uri")); + goto done; + } + if (uri_out) + uri = uri_out; /* Did domainMigratePrepare change URI? */ + assert (uri != NULL); + + /* Perform the migration. The driver isn't supposed to return + * until the migration is complete. + */ + if (domain->conn->driver->domainMigratePerform + (domain, cookie, cookielen, uri, flags, dname, bandwidth) == -1) + goto done; + + /* Get the destination domain and return it or error. + * 'domain' no longer actually exists at this point + * (or so we hope), but we still use the object in memory + * in order to get the name. + */ + dname = dname ? dname : domain->name; + if (dconn->driver->domainMigrateFinish) + ddomain = dconn->driver->domainMigrateFinish + (dconn, dname, cookie, cookielen, uri, flags); + else + ddomain = virDomainLookupByName (dconn, dname); + + done: + VIR_FREE (uri_out); + VIR_FREE (cookie); + return ddomain; +} + +static virDomainPtr +migrate_version_2 (virDomainPtr domain, + virConnectPtr dconn, + unsigned long flags, + const char *dname, + const char *uri, + unsigned long bandwidth) +{ + virDomainPtr ddomain = NULL; + char *uri_out = NULL; + char *cookie = NULL; + char *dom_xml = NULL; + int cookielen = 0, ret; + + /* Prepare the migration. + * + * The destination host may return a cookie, or leave cookie as + * NULL. + * + * The destination host MUST set uri_out if uri_in is NULL. + * + * If uri_in is non-NULL, then the destination host may modify + * the URI by setting uri_out. If it does not wish to modify + * the URI, it should leave uri_out as NULL. + */ + + /* In version 2 of the protocol, the prepare step is slightly + * different. We fetch the domain XML of the source domain + * and pass it to Prepare2. + */ + if (!domain->conn->driver->domainDumpXML) { + virLibConnError (domain->conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__); + return NULL; + } + dom_xml = domain->conn->driver->domainDumpXML (domain, + VIR_DOMAIN_XML_SECURE); + if (!dom_xml) + return NULL; + + ret = dconn->driver->domainMigratePrepare2 + (dconn, &cookie, &cookielen, uri, &uri_out, flags, dname, + bandwidth, dom_xml); + VIR_FREE (dom_xml); + if (ret == -1) + goto done; + + if (uri == NULL && uri_out == NULL) { + virLibConnError (domain->conn, VIR_ERR_INTERNAL_ERROR, + _("domainMigratePrepare2 did not set uri")); + goto done; + } + if (uri_out) + uri = uri_out; /* Did domainMigratePrepare2 change URI? */ + assert (uri != NULL); + + /* Perform the migration. The driver isn't supposed to return + * until the migration is complete. + */ + ret = domain->conn->driver->domainMigratePerform + (domain, cookie, cookielen, uri, flags, dname, bandwidth); + if (ret == -1) + goto done; + + /* In version 2 of the migration protocol, we pass the + * status code from the sender to the destination host, + * so it can do any cleanup if the migration failed. + */ + dname = dname ? dname : domain->name; + ddomain = dconn->driver->domainMigrateFinish2 + (dconn, dname, cookie, cookielen, uri, flags, ret); + + done: + VIR_FREE (uri_out); + VIR_FREE (cookie); + return ddomain; +} + /** * virDomainMigrate: * @domain: a domain object @@ -2930,140 +3070,52 @@ virDomainMigrate (virDomainPtr domain, const char *uri, unsigned long bandwidth) { - virConnectPtr conn; virDomainPtr ddomain = NULL; - char *uri_out = NULL; - char *cookie = NULL; - char *dom_xml = NULL; - int cookielen = 0, ret, version = 0; DEBUG("domain=%p, dconn=%p, flags=%lu, dname=%s, uri=%s, bandwidth=%lu", domain, dconn, flags, NULLSTR(dname), NULLSTR(uri), bandwidth); virResetLastError(); + /* First checkout the source */ if (!VIR_IS_CONNECTED_DOMAIN (domain)) { virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__); return NULL; } - conn = domain->conn; /* Source connection. */ - if (!VIR_IS_CONNECT (dconn)) { - virLibConnError (conn, VIR_ERR_INVALID_CONN, __FUNCTION__); + if (domain->conn->flags & VIR_CONNECT_RO) { + virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__); goto error; } - if (domain->conn->flags & VIR_CONNECT_RO) { - virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__); + /* Now checkout the destination */ + if (!VIR_IS_CONNECT (dconn)) { + virLibConnError (domain->conn, VIR_ERR_INVALID_CONN, __FUNCTION__); goto error; } if (dconn->flags & VIR_CONNECT_RO) { - /* NB, delibrately report error against source object, not dest here */ - virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__); + /* NB, deliberately report error against source object, not dest */ + virLibDomainError (domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__); goto error; } /* Check that migration is supported by both drivers. */ - if (VIR_DRV_SUPPORTS_FEATURE (conn->driver, conn, + if (VIR_DRV_SUPPORTS_FEATURE (domain->conn->driver, domain->conn, VIR_DRV_FEATURE_MIGRATION_V1) && VIR_DRV_SUPPORTS_FEATURE (dconn->driver, dconn, VIR_DRV_FEATURE_MIGRATION_V1)) - version = 1; - else if (VIR_DRV_SUPPORTS_FEATURE (conn->driver, conn, + ddomain = migrate_version_1 (domain, dconn, flags, dname, uri, bandwidth); + else if (VIR_DRV_SUPPORTS_FEATURE (domain->conn->driver, domain->conn, VIR_DRV_FEATURE_MIGRATION_V2) && VIR_DRV_SUPPORTS_FEATURE (dconn->driver, dconn, VIR_DRV_FEATURE_MIGRATION_V2)) - version = 2; + ddomain = migrate_version_2 (domain, dconn, flags, dname, uri, bandwidth); else { - virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__); + virLibConnError (domain->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__); goto error; } - /* Prepare the migration. - * - * The destination host may return a cookie, or leave cookie as - * NULL. - * - * The destination host MUST set uri_out if uri_in is NULL. - * - * If uri_in is non-NULL, then the destination host may modify - * the URI by setting uri_out. If it does not wish to modify - * the URI, it should leave uri_out as NULL. - */ - if (version == 1) { - ret = dconn->driver->domainMigratePrepare - (dconn, &cookie, &cookielen, uri, &uri_out, flags, dname, - bandwidth); - if (ret == -1) goto done; - if (uri == NULL && uri_out == NULL) { - virLibConnError (conn, VIR_ERR_INTERNAL_ERROR, - _("domainMigratePrepare did not set uri")); - goto done; - } - if (uri_out) uri = uri_out; /* Did domainMigratePrepare change URI? */ - - assert (uri != NULL); - } - else /* if (version == 2) */ { - /* In version 2 of the protocol, the prepare step is slightly - * different. We fetch the domain XML of the source domain - * and pass it to Prepare2. - */ - if (!conn->driver->domainDumpXML) { - virLibConnError (conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__); - goto error; - } - dom_xml = conn->driver->domainDumpXML (domain, - VIR_DOMAIN_XML_SECURE); - - if (!dom_xml) - goto error; + if (ddomain == NULL) + goto error; - ret = dconn->driver->domainMigratePrepare2 - (dconn, &cookie, &cookielen, uri, &uri_out, flags, dname, - bandwidth, dom_xml); - VIR_FREE (dom_xml); - if (ret == -1) goto done; - if (uri == NULL && uri_out == NULL) { - virLibConnError (conn, VIR_ERR_INTERNAL_ERROR, - _("domainMigratePrepare2 did not set uri")); - goto done; - } - if (uri_out) uri = uri_out; /* Did domainMigratePrepare2 change URI? */ - - assert (uri != NULL); - } - - /* Perform the migration. The driver isn't supposed to return - * until the migration is complete. - */ - ret = conn->driver->domainMigratePerform - (domain, cookie, cookielen, uri, flags, dname, bandwidth); - - if (version == 1) { - if (ret == -1) goto done; - /* Get the destination domain and return it or error. - * 'domain' no longer actually exists at this point - * (or so we hope), but we still use the object in memory - * in order to get the name. - */ - dname = dname ? dname : domain->name; - if (dconn->driver->domainMigrateFinish) - ddomain = dconn->driver->domainMigrateFinish - (dconn, dname, cookie, cookielen, uri, flags); - else - ddomain = virDomainLookupByName (dconn, dname); - } else /* if (version == 2) */ { - /* In version 2 of the migration protocol, we pass the - * status code from the sender to the destination host, - * so it can do any cleanup if the migration failed. - */ - dname = dname ? dname : domain->name; - ddomain = dconn->driver->domainMigrateFinish2 - (dconn, dname, cookie, cookielen, uri, flags, ret); - } - - done: - VIR_FREE (uri_out); - VIR_FREE (cookie); return ddomain; error: -- 1.6.0.6

2 2

[libvirt] [PATCH] Fix PCIe FLR detection
by Mark McLoughlin 31 Jul '09

31 Jul '09

PCIe DevCap register is actually 32 bits, not 16 bits. Since FLR is bit 28, we clearly are failing to detect FLR support. Known to fix device reset with some SR-IOV devices. * src/pci.c: fix pciDetectFunctionLevelReset() --- src/pci.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/pci.c b/src/pci.c index 4030a14..2dc2e1c 100644 --- a/src/pci.c +++ b/src/pci.c @@ -321,7 +321,7 @@ pciFindCapabilityOffset(pciDevice *dev, unsigned capability) static unsigned pciDetectFunctionLevelReset(pciDevice *dev) { - uint16_t caps; + uint32_t caps; uint8_t pos; /* The PCIe Function Level Reset capability allows @@ -331,7 +331,7 @@ pciDetectFunctionLevelReset(pciDevice *dev) * on SR-IOV NICs at the moment. */ if (dev->pcie_cap_pos) { - caps = pciRead16(dev, dev->pcie_cap_pos + PCI_EXP_DEVCAP); + caps = pciRead32(dev, dev->pcie_cap_pos + PCI_EXP_DEVCAP); if (caps & PCI_EXP_DEVCAP_FLR) { VIR_DEBUG("%s %s: detected PCIe FLR capability", dev->id, dev->name); return 1; -- 1.6.2.5

2 1

[libvirt] PATCH: Re-enable ESX driver on Mingw32
by Daniel P. Berrange 31 Jul '09

31 Jul '09

After finding an RPM of mingw32-curl for Fedora 11, I was able to successfully build the ESX driver on Ming32 with only a couple of tweaks. Daniel commit 8b135962ef026ef3ef1b5c5163b1213eeba9ec36 Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Thu Jul 30 16:40:29 2009 +0100 Enable ESX driver build on Mingw32 * autobuild.sh, mingw32-libvirt.spec.in: Enable esx on mingw32 * src/esx/esx_driver.c: Define AI_ADDRCONFIG if not set * src/esx/esx_util.c, src/esx/esx_vi_types.c: Always use %lld & friends, since gnulib guarentees we have these and not the target's own variants diff --git a/autobuild.sh b/autobuild.sh index 239372c..d3934ea 100755 --- a/autobuild.sh +++ b/autobuild.sh @@ -78,7 +78,6 @@ if [ -x /usr/bin/i686-pc-mingw32-gcc ]; then --without-openvz \ --without-one \ --without-phyp \ - --without-esx \ --without-netcf \ --without-libvirtd diff --git a/mingw32-libvirt.spec.in b/mingw32-libvirt.spec.in index 8fd08a6..40a8124 100644 --- a/mingw32-libvirt.spec.in +++ b/mingw32-libvirt.spec.in @@ -54,7 +54,6 @@ MinGW Windows libvirt virtualization library. --without-openvz \ --without-one \ --without-phyp \ - --without-esx \ --without-netcf \ --without-libvirtd make diff --git a/src/esx/esx_driver.c b/src/esx/esx_driver.c index e3d60a5..08d82f9 100644 --- a/src/esx/esx_driver.c +++ b/src/esx/esx_driver.c @@ -1769,7 +1769,7 @@ esxDomainGetInfo(virDomainPtr domain, virDomainInfoPtr info) for (value = perfMetricIntSeries->value; value != NULL; value = value->_next) { - VIR_DEBUG("value %"PRIi64, value->value); + VIR_DEBUG("value %lld", (long long int)value->value); } } } @@ -2399,8 +2399,8 @@ esxDomainGetSchedulerParameters(virDomainPtr domain, default: ESX_ERROR(domain->conn, VIR_ERR_INTERNAL_ERROR, - "Shares level has unknown value %"PRIi32, - sharesInfo->level); + "Shares level has unknown value %d", + (int)sharesInfo->level); goto failure; } diff --git a/src/esx/esx_util.c b/src/esx/esx_util.c index 5c87ea3..a113fde 100644 --- a/src/esx/esx_util.c +++ b/src/esx/esx_util.c @@ -41,6 +41,10 @@ virReportErrorHelper (conn, VIR_FROM_ESX, code, __FILE__, __FUNCTION__, \ __LINE__, fmt) +/* AI_ADDRCONFIG is missing on some systems. */ +#ifndef AI_ADDRCONFIG +# define AI_ADDRCONFIG 0 +#endif char * @@ -435,12 +439,12 @@ esxUtil_GetConfigLong(virConnectPtr conn, virConfPtr conf, const char *name, int esxUtil_GetConfigBoolean(virConnectPtr conn, virConfPtr conf, - const char *name, int *boolean, int default_, + const char *name, int *boolval, int default_, int optional) { virConfValuePtr value; - *boolean = default_; + *boolval = default_; value = virConfGetValue(conf, name); if (value == NULL) { @@ -465,9 +469,9 @@ esxUtil_GetConfigBoolean(virConnectPtr conn, virConfPtr conf, } if (STRCASEEQ(value->str, "true")) { - *boolean = 1; + *boolval = 1; } else if (STRCASEEQ(value->str, "false")) { - *boolean = 0; + *boolval = 0; } else { ESX_ERROR(conn, VIR_ERR_INTERNAL_ERROR, "Config entry '%s' must represent a boolean value " diff --git a/src/esx/esx_vi_types.c b/src/esx/esx_vi_types.c index 684b8c7..ab17460 100644 --- a/src/esx/esx_vi_types.c +++ b/src/esx/esx_vi_types.c @@ -962,7 +962,7 @@ esxVI_Int_DeepCopy(virConnectPtr conn, esxVI_Int **dest, esxVI_Int *src) /* esxVI_Int_Serialize */ ESX_VI__TEMPLATE__SERIALIZE_EXTRA(Int, "xsd:int", { - virBufferVSprintf(output, "%"PRIi32, item->value); + virBufferVSprintf(output, "%d", (int)item->value); }); /* esxVI_Int_SerializeList */ @@ -992,7 +992,7 @@ ESX_VI__TEMPLATE__LIST__APPEND(Long); /* esxVI_Long_Serialize */ ESX_VI__TEMPLATE__SERIALIZE_EXTRA(Long, "xsd:long", { - virBufferVSprintf(output, "%"PRIi64, item->value); + virBufferVSprintf(output, "%lld", (long long int)item->value); }); /* esxVI_Long_SerializeList */ -- |: 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 :|

2 1

[libvirt] PATCH: Fix permissions problem starting QEMU
by Daniel P. Berrange 31 Jul '09

31 Jul '09

There is a minor bug when running QEMU non-root, and having capng enabled. libvirt is unable to write the PID file in /var/run/libvirt/qemu, since its now owned by 'qemu', but libvirtd has dropped all capabilties at this point. The fix is to delay dropping capabilities until after the PID file has been created. We should also be sure to kill the child if writing the PID file fails Daniel commit ce246587178bc6539a3ea6181cdf06ea45878fd3 Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Thu Jul 30 14:58:16 2009 +0100 Fix problem writing QEMU pidfile * src/util.c: Don't drop capabilities until after the PID file has been written. Kill off child if writing the PID file fails * src/qemu_driver.c: Remove bogus trailing '/' in state dir diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 9fb8506..26897d3 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -468,7 +468,7 @@ qemudStartup(int privileged) { goto out_of_memory; if (virAsprintf(&qemu_driver->stateDir, - "%s/run/libvirt/qemu/", LOCAL_STATE_DIR) == -1) + "%s/run/libvirt/qemu", LOCAL_STATE_DIR) == -1) goto out_of_memory; } else { uid_t uid = geteuid(); diff --git a/src/util.c b/src/util.c index ee64b28..39aae24 100644 --- a/src/util.c +++ b/src/util.c @@ -513,12 +513,6 @@ __virExec(virConnectPtr conn, if ((hook)(data) != 0) _exit(1); - /* The hook above may need todo something privileged, so - * we delay clearing capabilities until now */ - if ((flags & VIR_EXEC_CLEAR_CAPS) && - virClearCapabilities() < 0) - _exit(1); - /* Daemonize as late as possible, so the parent process can detect * the above errors with wait* */ if (flags & VIR_EXEC_DAEMON) { @@ -543,6 +537,9 @@ __virExec(virConnectPtr conn, if (pid > 0) { if (pidfile && virFileWritePidPath(pidfile,pid)) { + kill(pid, SIGTERM); + usleep(500*1000); + kill(pid, SIGTERM); virReportSystemError(conn, errno, "%s", _("could not write pidfile")); _exit(1); @@ -551,6 +548,12 @@ __virExec(virConnectPtr conn, } } + /* The steps above may need todo something privileged, so + * we delay clearing capabilities until the last minute */ + if ((flags & VIR_EXEC_CLEAR_CAPS) && + virClearCapabilities() < 0) + _exit(1); + if (envp) execve(argv[0], (char **) argv, (char**)envp); else -- |: 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 :|

3 4

[libvirt] PATCH: Disable IPv6 on virtual network bridges
by Daniel P. Berrange 30 Jul '09

30 Jul '09

This is to address: https://bugzilla.redhat.com/show_bug.cgi?id=501934 which allows the guest to DOS the host IPv6 connectivity Daniel commit 763cf06ff76b4ded03a9b577cd8c541729190edc Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Thu Jul 30 16:34:56 2009 +0100 Disable IPv6 on virtual networks If the bridge device is configured to have IPv6 address and accept router advertisments, then a malicious guest can send out bogus advertisments and hijack/DOS host IPv6 connectivity * src/network_driver.c: Set accept_ra=0, disable_ipv6=1, autoconf=0 for IPv6 sysctl on virual network bridge devices diff --git a/src/network_driver.c b/src/network_driver.c index 1683631..eaea454 100644 --- a/src/network_driver.c +++ b/src/network_driver.c @@ -788,6 +788,55 @@ networkEnableIpForwarding(void) return virFileWriteStr("/proc/sys/net/ipv4/ip_forward", "1\n"); } +#define SYSCTL_PATH "/proc/sys" + +static int networkDisableIPV6(virConnectPtr conn, + virNetworkObjPtr network) +{ + char *field = NULL; + int ret = -1; + + if (virAsprintf(&field, SYSCTL_PATH "/net/ipv6/conf/%s/disable_ipv6", network->def->bridge) < 0) { + virReportOOMError(conn); + goto cleanup; + } + + if (virFileWriteStr(field, "1") < 0) { + virReportSystemError(conn, errno, + _("cannot enable %s"), field); + goto cleanup; + } + VIR_FREE(field); + + if (virAsprintf(&field, SYSCTL_PATH "/net/ipv6/conf/%s/accept_ra", network->def->bridge) < 0) { + virReportOOMError(conn); + goto cleanup; + } + + if (virFileWriteStr(field, "0") < 0) { + virReportSystemError(conn, errno, + _("cannot disable %s"), field); + goto cleanup; + } + VIR_FREE(field); + + if (virAsprintf(&field, SYSCTL_PATH "/net/ipv6/conf/%s/autoconf", network->def->bridge) < 0) { + virReportOOMError(conn); + goto cleanup; + } + + if (virFileWriteStr(field, "1") < 0) { + virReportSystemError(conn, errno, + _("cannot enable %s"), field); + goto cleanup; + } + + ret = 0; +cleanup: + VIR_FREE(field); + return ret; +} + static int networkStartNetworkDaemon(virConnectPtr conn, struct network_driver *driver, virNetworkObjPtr network) { @@ -806,6 +855,9 @@ static int networkStartNetworkDaemon(virConnectPtr conn, return -1; } + if (networkDisableIPV6(conn, network) < 0) + goto err_delbr; + if (brSetForwardDelay(driver->brctl, network->def->bridge, network->def->delay) < 0) goto err_delbr; -- |: 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 :|

2 4

[libvirt] [PATCH] BZ481602: Attempted to reproduce the bug led to issues around type mapping. They are resolved, and the code from the bug runs aginst hte test driver. Once published I will have the submitter reproduce it.
by Bryan Kearney 30 Jul '09

30 Jul '09

--- src/main/java/org/libvirt/Domain.java | 4 +--- src/main/java/org/libvirt/SchedLongParameter.java | 2 +- src/main/java/org/libvirt/SchedParameter.java | 2 ++ src/main/java/org/libvirt/SchedUintParameter.java | 2 +- src/main/java/org/libvirt/jna/Libvirt.java | 2 +- src/test/java/test.java | 8 ++++++++ 6 files changed, 14 insertions(+), 6 deletions(-) diff --git a/src/main/java/org/libvirt/Domain.java b/src/main/java/org/libvirt/Domain.java index a9a6b25..a7b49ee 100644 --- a/src/main/java/org/libvirt/Domain.java +++ b/src/main/java/org/libvirt/Domain.java @@ -607,13 +607,11 @@ public class Domain { * @throws LibvirtException */ public void setSchedulerParameters(SchedParameter[] params) throws LibvirtException { - IntByReference nParams = new IntByReference(); - nParams.setValue(params.length); virSchedParameter[] input = new virSchedParameter[params.length]; for (int x = 0; x < params.length; x++) { input[x] = SchedParameter.toNative(params[x]); } - libvirt.virDomainSetSchedulerParameters(VDP, input, nParams); + libvirt.virDomainSetSchedulerParameters(VDP, input, params.length); processError(); } diff --git a/src/main/java/org/libvirt/SchedLongParameter.java b/src/main/java/org/libvirt/SchedLongParameter.java index 50628b0..7999777 100644 --- a/src/main/java/org/libvirt/SchedLongParameter.java +++ b/src/main/java/org/libvirt/SchedLongParameter.java @@ -21,7 +21,7 @@ public final class SchedLongParameter extends SchedParameter { } public int getType() { - return 2; + return 3; } public String getTypeAsString() { diff --git a/src/main/java/org/libvirt/SchedParameter.java b/src/main/java/org/libvirt/SchedParameter.java index 92b3547..370b0da 100644 --- a/src/main/java/org/libvirt/SchedParameter.java +++ b/src/main/java/org/libvirt/SchedParameter.java @@ -4,6 +4,7 @@ import java.util.Arrays; import org.libvirt.jna.Libvirt; import org.libvirt.jna.virSchedParameter; +import org.libvirt.jna.virSchedParameterValue; import com.sun.jna.Native; @@ -43,6 +44,7 @@ public abstract class SchedParameter { public static virSchedParameter toNative(SchedParameter param) { virSchedParameter returnValue = new virSchedParameter(); + returnValue.value = new virSchedParameterValue() ; returnValue.field = Arrays.copyOf(param.field.getBytes(), Libvirt.VIR_DOMAIN_SCHED_FIELD_LENGTH); returnValue.type = param.getType(); switch (param.getType()) { diff --git a/src/main/java/org/libvirt/SchedUintParameter.java b/src/main/java/org/libvirt/SchedUintParameter.java index 1b96758..9305579 100644 --- a/src/main/java/org/libvirt/SchedUintParameter.java +++ b/src/main/java/org/libvirt/SchedUintParameter.java @@ -22,7 +22,7 @@ public final class SchedUintParameter extends SchedParameter { } public int getType() { - return 3; + return 2; } public String getTypeAsString() { diff --git a/src/main/java/org/libvirt/jna/Libvirt.java b/src/main/java/org/libvirt/jna/Libvirt.java index 62f7ba3..0b30656 100644 --- a/src/main/java/org/libvirt/jna/Libvirt.java +++ b/src/main/java/org/libvirt/jna/Libvirt.java @@ -107,7 +107,7 @@ public interface Libvirt extends Library { public int virDomainSetAutostart(DomainPointer virDomainPtr, int autoStart); public int virDomainSetMaxMemory(DomainPointer virDomainPtr, NativeLong maxMemory); public int virDomainSetMemory(DomainPointer virDomainPtr, NativeLong maxMemory); - public int virDomainSetSchedulerParameters(DomainPointer virDomainPtr, virSchedParameter[] params, IntByReference nparams); + public int virDomainSetSchedulerParameters(DomainPointer virDomainPtr, virSchedParameter[] params, int nparams); public int virDomainSetVcpus(DomainPointer virDomainPtr, int nvcpus); public int virDomainShutdown(DomainPointer virDomainPtr); public int virDomainSuspend(DomainPointer virDomainPtr); diff --git a/src/test/java/test.java b/src/test/java/test.java index 9427cc5..be12d5e 100644 --- a/src/test/java/test.java +++ b/src/test/java/test.java @@ -245,6 +245,14 @@ public class test { for(SchedParameter c: testDomain.getSchedulerParameters()){ System.out.println(c.getTypeAsString() +":"+ c.field +":"+ c.getValueAsString()); } + + // test setting a scheduled parameter + SchedUintParameter[] pars = new SchedUintParameter[1]; + pars[0] = new SchedUintParameter(); + pars[0].field = "weight"; + pars[0].value = 100; + testDomain.setSchedulerParameters(pars); + System.out.println("virDomainGetUUID:" + testDomain.getUUID()); for(int c: testDomain.getUUID()) System.out.print(String.format("%02x", c)); -- 1.6.0.6

1 0

[libvirt] Pre-0.7.0 release for testing
by Daniel Veillard 30 Jul '09

30 Jul '09

Considering the large amount of new code and features expected for 0.7.0 at the end of the week, I made a pre-release at ftp://libvirt.org/libvirt/libvirt-0.7.0-0.1.gitf055724.tar.gz if people have time to test and report, that would be great ! I expect to make the actual release this thursday, unless there is too many issues in which case I would postpone to beginning of next week (or whenever main problems get fixed...) thanks ! Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

2 4

[libvirt] Fix -Werror build error regarding to '#ifdef IFF_VNET_HDR'
by Maximilian Wilhelm 30 Jul '09

30 Jul '09

Hi! Our build bot found a build error when building libVirt with -Werror introduced in b4f62abbf1191c8fbab3306b4bf2f2567e18067f kvm/virtio: Set IFF_VNET_HDR when setting up tap fds The attached patch kind of fixes the build error but I'm not sure, if this is the final solution for this problem, though... Ciao Max -- The real problem with C++ for kernel modules is: the language just sucks. -- Linus Torvalds

2 3

[libvirt] [PATCH] Drop the leading lib from libvirt. Based on testing by thomas@scripty.at this is a better way to load the libarary
by Bryan Kearney 30 Jul '09

30 Jul '09

--- src/main/java/org/libvirt/jna/Libvirt.java | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/main/java/org/libvirt/jna/Libvirt.java b/src/main/java/org/libvirt/jna/Libvirt.java index 33914ac..62f7ba3 100644 --- a/src/main/java/org/libvirt/jna/Libvirt.java +++ b/src/main/java/org/libvirt/jna/Libvirt.java @@ -20,7 +20,7 @@ public interface Libvirt extends Library { public int authCallback(virConnectCredential cred, int ncred, Pointer cbdata); } - Libvirt INSTANCE = (Libvirt) Native.loadLibrary("libvirt", Libvirt.class); + Libvirt INSTANCE = (Libvirt) Native.loadLibrary("virt", Libvirt.class); // Constants we need public static int VIR_UUID_BUFLEN = 16; -- 1.6.0.6

3 2

[libvirt] RFC: libvirt java bindings based on JNA
by Bryan Kearney 30 Jul '09

30 Jul '09

I would like to get some comments on an initial cut of the java bindings on top of JNA[1]. They are not 100% complete (see below), but they are good enough to be criticized. In addition, if there are features additions to the API which people are interested in, I would appreciate the feedback on that. JNA is a java implementation of the Foreign Function Interface [2]. The main benefit is that it removes the need for all the nasty C glue code. What replaces it is pure java mapping files. If you look in the org.libvirt.jna package in the code, you will see that the 2928 lines of jni code have been replaced by 427 lines of java code. To be fair, some of the business logic from the C code is now in the org.libvirt package... but you get the point. The original java API has remained the same. The open issues with the code are: 1) Authorization callback has not been tested. 2) GetUUID is not returning the correct int array. 3) Find by UUID is not working These all are mapping issues to be worked out. The rest of the code should be working, but I would appreciate it being exercised. The main drawback to this approach is that it requires the use of java 1.6. I looked on EPEL, F10, and F11.. and java 1.6 and the JNA jar file are available so I did not believe this was an issue. I will send a patch stream to the list, but you can also get the code via an SRPM [3] or via github [4]. -- bk [1] https://jna.dev.java.net/ [2] http://en.wikipedia.org/wiki/Foreign_function_interface [3] http://bkearney.fedorapeople.org/libvirt-java-0.3.0pre-1.fc10.src.rpm [4] http://github.com/bkearney/libvirt-java/tree/jna

3 11

[libvirt] PATCH: Always start dnsmasq to provide DNS
by Daniel P. Berrange 30 Jul '09

30 Jul '09

Currently we only start dnsmasq if DHCP ranges are configured. https://bugzilla.redhat.com/show_bug.cgi?id=508965 points out that this prevents use of DNS in guests using static IP config. Thus this patch makes us always start dnsmasq, regardless of DHCP config if we have an IP addr configured. There is no need for the --no-dhcp-interface= that BZ mentions, since that's just to allow you to restrict which interfaces dnsmasq does DHCP on - we've already restricted it to just out interface, and leaving out the DHCP ranges disables its DHCP server for every interface Regards, Daniel commit deb71927a3551b3a1554351ef098644aec978a09 Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Thu Jul 30 14:52:31 2009 +0100 Allow dnsmasq to provide DNS without DHCP * src/network_driver.c: Always start dnsmasq to allow it to provide DNS, even if no DHCP ranges are enabled diff --git a/src/network_driver.c b/src/network_driver.c index 9621782..1683631 100644 --- a/src/network_driver.c +++ b/src/network_driver.c @@ -136,8 +136,9 @@ networkFindActiveConfigs(struct network_driver *driver) { brHasBridge(driver->brctl, obj->def->bridge) == 0) { obj->active = 1; - /* Finally try and read dnsmasq pid if any DHCP ranges are set */ - if (obj->def->nranges && + /* Finally try and read dnsmasq pid if any */ + if ((obj->def->ipAddress || + obj->def->nranges) && virFileReadPid(NETWORK_PID_DIR, obj->def->name, &obj->dnsmasqPid) == 0) { @@ -844,7 +845,8 @@ static int networkStartNetworkDaemon(virConnectPtr conn, goto err_delbr2; } - if (network->def->nranges && + if ((network->def->ipAddress || + network->def->nranges) && dhcpStartDhcpDaemon(conn, network) < 0) goto err_delbr2; -- |: 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 :|

2 1

[libvirt] [PATCH] Remove MAX_TAP_ID, take 3
by Aron Griffis 30 Jul '09

30 Jul '09

This is a resend of take 2 to fix formatting problems in the patch. No other changes. As far as I can tell, there's no reason to format the device string in brAddTap(). Delegate the job to TUNSETIFF, thereby removing the loop and the MAX_TAP_ID artificial limit. This patch allows me to get 421 guests running before hitting other limits. Signed-off-by: Aron Griffis <aron.griffis(a)hp.com> diff --git a/src/bridge.c b/src/bridge.c index 0509afd..ec37192 100644 --- a/src/bridge.c +++ b/src/bridge.c @@ -49,8 +49,6 @@ #include "util.h" #include "logging.h" -#define MAX_TAP_ID 256 - #define JIFFIES_TO_MS(j) (((j)*1000)/HZ) #define MS_TO_JIFFIES(ms) (((ms)*HZ)/1000) @@ -466,76 +464,52 @@ brAddTap(brControl *ctl, int vnet_hdr, int *tapfd) { - int id, subst, fd; + int fd, len; + struct ifreq ifr = {0}; if (!ctl || !ctl->fd || !bridge || !ifname) return EINVAL; - subst = id = 0; - - if (strstr(*ifname, "%d")) - subst = 1; - if ((fd = open("/dev/net/tun", O_RDWR)) < 0) return errno; - if (vnet_hdr) - vnet_hdr = brProbeVnetHdr(fd); + ifr.ifr_flags = IFF_TAP|IFF_NO_PI; - do { - struct ifreq try; - int len; +#ifdef IFF_VNET_HDR + if (vnet_hdr && brProbeVnetHdr(fd)) + ifr.ifr_flags |= IFF_VNET_HDR; +#endif - memset(&try, 0, sizeof(struct ifreq)); + strncpy(ifr.ifr_name, *ifname, IFNAMSIZ-1); - try.ifr_flags = IFF_TAP|IFF_NO_PI; + if (ioctl(fd, TUNSETIFF, &ifr) < 0) + goto error; -#ifdef IFF_VNET_HDR - if (vnet_hdr) - try.ifr_flags |= IFF_VNET_HDR; -#endif + len = strlen(ifr.ifr_name); + if (len >= BR_IFNAME_MAXLEN - 1) { + errno = EINVAL; + goto error; + } - if (subst) { - len = snprintf(try.ifr_name, BR_IFNAME_MAXLEN, *ifname, id); - if (len >= BR_IFNAME_MAXLEN) { - errno = EADDRINUSE; - goto error; - } - } else { - len = strlen(*ifname); - if (len >= BR_IFNAME_MAXLEN - 1) { - errno = EINVAL; - goto error; - } - - strncpy(try.ifr_name, *ifname, len); - try.ifr_name[len] = '\0'; - } - - if (ioctl(fd, TUNSETIFF, &try) == 0) { - /* We need to set the interface MTU before adding it - * to the bridge, because the bridge will have its - * MTU adjusted automatically when we add the new interface. - */ - if ((errno = brSetInterfaceMtu(ctl, bridge, try.ifr_name))) - goto error; - if ((errno = brAddInterface(ctl, bridge, try.ifr_name))) - goto error; - if ((errno = brSetInterfaceUp(ctl, try.ifr_name, 1))) - goto error; - if (!tapfd && - (errno = ioctl(fd, TUNSETPERSIST, 1))) - goto error; - VIR_FREE(*ifname); - if (!(*ifname = strdup(try.ifr_name))) - goto error; - if (tapfd) - *tapfd = fd; - return 0; - } - - id++; - } while (subst && id <= MAX_TAP_ID); + /* We need to set the interface MTU before adding it + * to the bridge, because the bridge will have its + * MTU adjusted automatically when we add the new interface. + */ + if ((errno = brSetInterfaceMtu(ctl, bridge, ifr.ifr_name))) + goto error; + if ((errno = brAddInterface(ctl, bridge, ifr.ifr_name))) + goto error; + if ((errno = brSetInterfaceUp(ctl, ifr.ifr_name, 1))) + goto error; + if (!tapfd && + (errno = ioctl(fd, TUNSETPERSIST, 1))) + goto error; + VIR_FREE(*ifname); + if (!(*ifname = strdup(ifr.ifr_name))) + goto error; + if (tapfd) + *tapfd = fd; + return 0; error: close(fd);

4 9

[libvirt] VM creation failed : Permission denied : bind(unix:/var/run/libvirt/qemu//VM.monitor)
by Pierre-Gilles Mialon 30 Jul '09

30 Jul '09

I use Ovirt under Fedora 11 with the latest developpement package and the VM creation failed with : bind(unix:/var/run/libvirt/qemu//VM-4.monitor): Permission denied See the logs : On the Ovirt Node ==> /var/log/messages <== Jul 29 20:43:37 node62 kernel: device vnet0 entered promiscuous mode Jul 29 20:43:37 node62 kernel: breth0: port 2(vnet0) entering learning state Jul 29 20:43:37 node62 libvirtd: 20:43:37.873: info : qemudDispatchSignalEvent:370 : Received unexpected signal 17 Jul 29 20:43:37 node62 libvirtd: 20:43:37.877: info : qemudDispatchSignalEvent:370 : Received unexpected signal 17 Jul 29 20:43:37 node62 kernel: breth0: port 2(vnet0) entering disabled state Jul 29 20:43:37 node62 kernel: device vnet0 left promiscuous mode Jul 29 20:43:37 node62 kernel: breth0: port 2(vnet0) entering disabled state Jul 29 20:43:37 node62 libvirtd: 20:43:37.977: error : qemudReadLogOutput:816 : internal error Process exited while reading console log output Jul 29 20:43:37 node62 libvirtd: 20:43:37.977: error : qemudWaitForMonitor:1103 : internal error unable to start guest: bind(unix:/var/run/libvirt/qemu//VM-4.monitor): Permission denied#012qemu: could not open monitor device 'unix:/var/run/libvirt/qemu//VM-4.monitor,server,nowait'#012 Jul 29 20:43:38 node62 libvirtd: 20:43:38.073: error : qemudDomainLookupByUUID:2644 : Domain not found: no domain with matching uuid '492f836f-5123-e185-39c2-09c5dd43a7f6' Jul 29 20:43:38 node62 libvirt-qpid: Error: virDomainLookupByUUIDString Subsystem qemu: Domain not found: no domain with matching uuid '492f836f-5123- e185-39c2-09c5dd43a7f6' in NodeWrap.cpp:syncDomains:241 code: 42 ==> /var/log/libvirt/qemu/VM-4.log <== LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin HOME=/root USER=root LOGNAME=root /usr/bin/qemu-kvm -S -M pc -m 256 -smp 1 -name VM-4 -uuid 492f836f-5123- e185-39c2-09c5dd43a7f6 -monitor unix:/var/run/libvirt/qemu//VM-4.monitor,server,nowait -boot n -net nic,macaddr=00:16:3e:10:de:fe,vlan=0,name=nic.0 -net tap,fd=18,vlan=0,name=tap.0 -serial pty -parallel none -usb -vnc 0.0.0.0:0 - vga cirrus bind(unix:/var/run/libvirt/qemu//VM-4.monitor): Permission denied qemu: could not open monitor device 'unix:/var/run/libvirt/qemu//VM-4.monitor,server,nowait' ==> # rpm -qa | grep libvirt <== libvirt-0.7.0-0.2.gitf055724.fc11.x86_64 libvirt-python-0.7.0-0.2.gitf055724.fc11.x86_64 libvirt-client-0.7.0-0.2.gitf055724.fc11.x86_64 libvirt-qpid-0.2.17-0.fc11.x86_64 Regards -- Pierre-Gilles Mialon Linagora :: http://www.linagora.com Responsable hébergement :: Head of Hosting services pmialon(a)linagora.com :: +33.1 58 18 65 46

5 11

[libvirt] VMware ESX: Scheduler documentation and cleanup
by Matthias Bolte 30 Jul '09

30 Jul '09

Add some documentation about the CPU scheduler parameters and remove some old, unnecessary compensation code, since virsh uses the proposed parameter types now. Documentation in HTML form will be added soon. Matthias

2 1

[libvirt] vcpupin broken towards Xen using xend
by Henrik Persson 30 Jul '09

30 Jul '09

Hi, During the update of the cpu affinity map format, the old format was broken ( '[' was switched with ']'). Also, the mapstr might contain garbage since the string is not initiate (however buf incorrectly is). This small patch fixes both the problems. Regards, Henrik

2 1

[libvirt] PATCH: Fix RPM upgrades to create qemu user
by Daniel P. Berrange 30 Jul '09

30 Jul '09

If upgrading from F11 to F12 it is possible the qemu user/group does not yet exist. Since the libvirt RPM needs to deploy files owned by this user/group, regardless of whether QEMU is yet installed, we need to make sure to create this user/group in %pre, otherwise fiels will end up with wrong ownership. Daniel diff -u -p -r1.159 libvirt.spec --- libvirt.spec 29 Jul 2009 14:58:18 -0000 1.159 +++ libvirt.spec 30 Jul 2009 11:06:53 -0000 @@ -419,6 +419,18 @@ chmod 0644 $RPM_BUILD_ROOT%{_sysconfdir} %clean rm -fr %{buildroot} +%pre +%if 0%{?fedora} >= 12 +# Normally 'setup' adds this in /etc/passwd, but this is +# here for case of upgrades from earlier Fedora. This +# UID/GID pair is reserved for qemu:qemu +getent group kvm >/dev/null || groupadd -g 36 -r kvm +getent group qemu >/dev/null || groupadd -g 107 -r qemu +getent passwd qemu >/dev/null || \ + useradd -r -u 107 -g qemu -G kvm -d / -s /sbin/nologin \ + -c "qemu user" qemu +%endif + %post %if %{with_libvirtd} -- |: 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 :|

2 1

[libvirt] Feature request? Console tunnelling..
by r-c＠rocketmail.com 30 Jul '09

30 Jul '09

Hello libvirt-list.. Firstly I hope this is the appropriate forum for this question, I tried the IRC channel but didn't get any response... Anyway, currently libvirt has console functionality built into it that can be used on a Dom0 e.g. 'virsh console xenU' to connect and open guest xenU's console. I think it would be excellent if there could be a feature added to the libvirt daemon to operate this console externally! E.g. from an external machine I can do (in python) something like: domain.openConsole() .readConsole() .writeConsole() flushConsole()... hopefully that could be very nice to implement a remote Ajaxterm with more security and no reliance to run unnecessary services on the Dom0 since libvirtd is already running :-) I hope this makes sense, but if you perhaps thinks I going about this the wrong way please let me know! Cheers

2 1

Re: [libvirt] VMware support and Server 2.0.1
by Matthias Bolte 30 Jul '09

30 Jul '09

2009/7/29 Timo Makinen <tmakinen(a)ee.oulu.fi>: > On Wed, Jul 29, 2009 at 06:12:59PM +0200, Matthias Bolte wrote: >> 2009/7/27 Timo Makinen <tmakinen(a)ee.oulu.fi>: [...] > The dumpxml seems to be failing with this config option: > > ethernet0.connectionType = "custom" > > without this it works correctly. Ah, I found the cause. ethernet0.connectionType = "custom" was a good hint. I missed to initialize one variable correctly that is only used in the ethernet0.connectionType = "custom" case. >> Could you do some basic testing beside listing and dumpxml, as I have >> currently no VMware GSX installation at hand? Just test if the >> following commands work as expected in virsh: >> >> - start >> - shutdown >> - reboot >> - suspend >> - resume >> - nodeinfo >> - dominfo >> >> Regards, >> Matthias > > All these worked out of the box. Nice! Could you test the attached patch? It adds handling for the port in the URI, extends the version checking for GSX 2.0 and fixes the bug in VMX parsing. Matthias

2 1

[libvirt] [PATCH] Modify the java bindings html page to reflect the new repositories, JNA, and maven access
by Bryan Kearney 29 Jul '09

29 Jul '09

--- docs/java.html.in | 109 +++++++++++++++++++++++++++-------------------------- 1 files changed, 55 insertions(+), 54 deletions(-) diff --git a/docs/java.html.in b/docs/java.html.in index 2d9a24f..b0db2e2 100644 --- a/docs/java.html.in +++ b/docs/java.html.in @@ -4,10 +4,9 @@ <h1>Java API bindings</h1> <h2>Presentation</h2> - <p>The Java bindings are currently a work in progress based mostly -on the work of Toth Istvan. The first usable release is 0.2.0, where -most of the naming conventions were defined. Further release will try -as much as possible to stay compatible</p> + <p>The Java bindings make use of <a href="https://jna.dev.java.net/">JNA</a> + to expose the C API in a Java friendly way. The bindings are are based on + work initiated by Toth Istvan.</p> <h2>Getting it</h2> <p> @@ -19,82 +18,84 @@ as much as possible to stay compatible</p> <li><a href="http://libvirt.org/sources/java/">libvirt.org HTTP server</a></li> </ul> -<h2>Hourly development snapshots</h2> -<p> Once an hour, an automated snapshot is made from the latest CVS server - source tree. These snapshots should be usable, but we make no guarantees - about their stability: </p> +<h3>Maven</h3> +<p>A maven repository is located at <a href="http://www.libvirt.org/maven2/">http://www.libvirt.org/maven2/</a> +which you can use to include this in your maven projects.</p> -<ul> - <li><a href="ftp://libvirt.org/libvirt/java/libvirt-java-snapshot.tar.gz">libvirt.org FTP server</a></li> - <li><a href="http://libvirt.org/sources/java/libvirt-java-snapshot.tar.gz">libvirt.org HTTP server</a></li> -</ul> +<h2>GIT source repository</h2> +<p> The Java bindings code source is now maintained in a <a +href="http://git-scm.com/">git</a> repository available on +<a href="http://libvirt.org/git/">libvirt.org</a>: +</p> +<pre> +git clone git://libvirt.org/libvirt-java.git +</pre> +<p> +It can also be browsed at +</p> +<pre> -<h2>CVS repository access</h2> +<a href="http://libvirt.org/git/?p=libvirt-java.git;a=summary">http://libvirt.org/git/?p=libvirt-java.git;a=summary</a> +</pre> -<p> The master source repository uses <a -href="http://ximbiot.com/cvs/cvshome/docs/">CVS</a> and anonymous access -is provided. Prior to accessing the server is it necessary to authenticate -using the password <code>anoncvs</code>. This can be accomplished with -the <code>cvs login</code> command: +<p></p> +<h2>GIT repository mirror</h2> + +<p> + The source code is also mirrored on <a href="http://www.gitorious.org">gitorious</a> + to aid in public development. You clone the repo with </p> -<pre> - # cvs -d :pserver:anoncvs@libvirt.org:2401/data/cvs login +<pre> + git clone git://gitorious.org/libvirt/libvirt-java.git </pre> <p> - Once authenticated, a checkout can be obtained using + It can also be browsed at </p> <pre> - # cvs -d :pserver:anoncvs@libvirt.org:2401/data/cvs co libvirt-java + <a href="http://gitorious.org/libvirt/libvirt-java">http://gitorious.org/libvirt/libvirt-java</a> </pre> -<p> -The libvirt-java build process uses GNU autotools, so after obtaining a checkout -it is necessary to generate the configure script and Makefile.in templates -using the <code>autogen.sh</code> command. As an example, to do a complete -build and install it into your home directory run: +<p></p> +<h2>CVS repository access (Deprecated)</h2> + +<p> The old source repository uses <a +href="http://ximbiot.com/cvs/cvshome/docs/">CVS</a> and anonymous access +is provided. Prior to accessing the server is it necessary to authenticate +using the password <code>anoncvs</code>. This can be accomplished with +the <code>cvs login</code> command: </p> + <pre> - ./autogen.sh --prefix=$HOME/usr - make - make install + # cvs -d :pserver:anoncvs@libvirt.org:2401/data/cvs login </pre> -<p>Other build system are not currently available but as usual we take patches -(Eclipse/ant/...) would be welcome !</p> -<p> currently libvirt-java requires a version of libvirt >= 0.4.0 to be -installed as well as the associated libvirt development files/package. It -should compile with any java implementation >= 1.5, the bindings uses -the enum construct which appeared only in that version.</p> -<p>Libvirt-java bindings development occurs on the same mailing-list as -the normal libvirt work see <a href="contact.html">the associated contact -page</a></p> - -<h2>GIT repository mirror</h2> <p> - The CVS source repository is also mirrored using GIT, and is available - for anonymous access via: + Once authenticated, a checkout can be obtained using </p> -<pre> - git clone git://git.et.redhat.com/libvirt-java +<pre> + # cvs -d :pserver:anoncvs@libvirt.org:2401/data/cvs co libvirt-java </pre> -<p> - It can also be browsed at -</p> +<p></p> +<h2>Building</h2> +<p>The code is built using ant, and assumes that you have the jna jar installed. Once you have downloaded +the code you can build the code with</p> <pre> - <a href="http://git.et.redhat.com/?p=libvirt-java.git;a=summary">http://git.et.redhat.com/?p=libvirt-java.git;a=summary</a> +% cd libvirt-java +% ant build </pre> + <h2>Content</h2> <p>The bindings are articulated around a few -classes in the <code>org/libvirt</code> namespace, notably the +classes in the <code>org/libvirt</code> package, notably the <code>Connect</code>, <code>Domain</code> and <code>Network</code> ones. Functions in the <a href="html/libvirt-libvirt.html">C API</a> taking <code>virConnectPtr</code>, <code>virDomainPtr</code> or @@ -112,13 +113,13 @@ lower case, for example the C functions:</p> </p> <p>become</p> <p> - <code>virConn::numOfDomains()</code> + <code>virConn.numOfDomains()</code> </p> <p> - <code>virDomain::setMaxMemory(long memory)</code> + <code>virDomain.setMaxMemory(long memory)</code> </p> <p> There is of course some functions where the mapping is less direct -and using extra classes to map complex arguments. The <a href="http://libvirt.org/org/libvirt/package-summary.html">Javadoc</a> is available online or as +and using extra classes to map complex arguments. The <a href="http://libvirt.org/sources/java/javadoc">Javadoc</a> is available online or as part of a separate libvirt-java-javadoc package.</p> <p>So let's look at a simple example inspired from the <code>test.java</code> test found in <code>src</code> in the source tree:</p> @@ -147,7 +148,7 @@ public class minitest { <p>There is not much to comment about it, it really is a straight mapping from the C API, the only points to notice are:</p> <ul> - <li>the import of the modules in the <code><span style="color: #0071FF; background-color: #FFFFFF">org.libvirt</span></code> namespace</li> + <li>the import of the modules in the <code><span style="color: #0071FF; background-color: #FFFFFF">org.libvirt</span></code> package</li> <li>getting a connection to the hypervisor, in that case using the readonly access to the default test hypervisor.</li> <li>getting an object representing the test domain using <span style="color: #007F00; background-color: #FFFFFF">lookupByName</span></li> -- 1.6.0.6

2 1

[libvirt] ANNOUNCE: Java Libvirt bindings version 0.3.0
by Bryan Kearney 29 Jul '09

29 Jul '09

I am happy to announce the availability of version 0.3.0 of the Libvirt Java bindings. The main change in this release is the use of JNA [1] to access the Libvirt C code instead of JNI. This results in all the glue code being pure Java which should make it easier to keep these bindings in sync with the C code. The RPMS are in fedora-testing for 10, 11 and rawhide. They will move to updates per the normal process. The code can be directly downloaded from [2]. The updated repository has been moved to git, which can be found at [3]. I am able to generate a maven repository for the bits, and was planning on putting a maven repo on libvirt.org unless folks would prefer the bits to be mirrored in a different location. Please let me know. Thanks! -- bk [1] https://jna.dev.java.net/ [2] http://www.libvirt.org/sources/java/libvirt-java-0.3.0.tar.gz [3] http://www.libvirt.org/git/?p=libvirt-java.git;a=summary

1 0

[libvirt] [PATCH] Remove MAX_TAP_ID
by Aron Griffis 29 Jul '09

29 Jul '09

As far as I can tell, there's no reason to format the device string in brAddTap(). Delegate the job to TUNSETIFF, thereby removing the loop and the MAX_TAP_ID artificial limit. This patch allows me to get 421 guests running before hitting other limits. Signed-off-by: Aron Griffis <aron.griffis(a)hp.com> --- src/bridge.c | 94 +++++++++++++++++++++------------------------------------- 1 files changed, 34 insertions(+), 60 deletions(-) diff --git a/src/bridge.c b/src/bridge.c index 0509afd..ec37192 100644 --- a/src/bridge.c +++ b/src/bridge.c @@ -49,8 +49,6 @@ #include "util.h" #include "logging.h" -#define MAX_TAP_ID 256 - #define JIFFIES_TO_MS(j) (((j)*1000)/HZ) #define MS_TO_JIFFIES(ms) (((ms)*HZ)/1000) @@ -466,76 +464,52 @@ brAddTap(brControl *ctl, int vnet_hdr, int *tapfd) { - int id, subst, fd; + int fd, len; + struct ifreq ifr = {0}; if (!ctl || !ctl->fd || !bridge || !ifname) return EINVAL; - subst = id = 0; - - if (strstr(*ifname, "%d")) - subst = 1; - if ((fd = open("/dev/net/tun", O_RDWR)) < 0) return errno; - if (vnet_hdr) - vnet_hdr = brProbeVnetHdr(fd); + ifr.ifr_flags = IFF_TAP|IFF_NO_PI; - do { - struct ifreq try; - int len; +#ifdef IFF_VNET_HDR + if (vnet_hdr && brProbeVnetHdr(fd)) + ifr.ifr_flags |= IFF_VNET_HDR; +#endif - memset(&try, 0, sizeof(struct ifreq)); + strncpy(ifr.ifr_name, *ifname, IFNAMSIZ-1); - try.ifr_flags = IFF_TAP|IFF_NO_PI; + if (ioctl(fd, TUNSETIFF, &ifr) < 0) + goto error; -#ifdef IFF_VNET_HDR - if (vnet_hdr) - try.ifr_flags |= IFF_VNET_HDR; -#endif + len = strlen(ifr.ifr_name); + if (len >= BR_IFNAME_MAXLEN - 1) { + errno = EINVAL; + goto error; + } - if (subst) { - len = snprintf(try.ifr_name, BR_IFNAME_MAXLEN, *ifname, id); - if (len >= BR_IFNAME_MAXLEN) { - errno = EADDRINUSE; - goto error; - } - } else { - len = strlen(*ifname); - if (len >= BR_IFNAME_MAXLEN - 1) { - errno = EINVAL; - goto error; - } - - strncpy(try.ifr_name, *ifname, len); - try.ifr_name[len] = '\0'; - } - - if (ioctl(fd, TUNSETIFF, &try) == 0) { - /* We need to set the interface MTU before adding it - * to the bridge, because the bridge will have its - * MTU adjusted automatically when we add the new interface. - */ - if ((errno = brSetInterfaceMtu(ctl, bridge, try.ifr_name))) - goto error; - if ((errno = brAddInterface(ctl, bridge, try.ifr_name))) - goto error; - if ((errno = brSetInterfaceUp(ctl, try.ifr_name, 1))) - goto error; - if (!tapfd && - (errno = ioctl(fd, TUNSETPERSIST, 1))) - goto error; - VIR_FREE(*ifname); - if (!(*ifname = strdup(try.ifr_name))) - goto error; - if (tapfd) - *tapfd = fd; - return 0; - } - - id++; - } while (subst && id <= MAX_TAP_ID); + /* We need to set the interface MTU before adding it + * to the bridge, because the bridge will have its + * MTU adjusted automatically when we add the new interface. + */ + if ((errno = brSetInterfaceMtu(ctl, bridge, ifr.ifr_name))) + goto error; + if ((errno = brAddInterface(ctl, bridge, ifr.ifr_name))) + goto error; + if ((errno = brSetInterfaceUp(ctl, ifr.ifr_name, 1))) + goto error; + if (!tapfd && + (errno = ioctl(fd, TUNSETPERSIST, 1))) + goto error; + VIR_FREE(*ifname); + if (!(*ifname = strdup(ifr.ifr_name))) + goto error; + if (tapfd) + *tapfd = fd; + return 0; error: close(fd);

2 2

[libvirt] [PATCH] Remove MAX_TAP_ID, take 2
by Aron Griffis 29 Jul '09

29 Jul '09

(Sorry, my first posting included some gnulib droppings in the patch. This removes that, everything else is the same.) As far as I can tell, there's no reason to format the device string in brAddTap(). Delegate the job to TUNSETIFF, thereby removing the loop and the MAX_TAP_ID artificial limit. This patch allows me to get 421 guests running before hitting other limits. Signed-off-by: Aron Griffis <aron.griffis(a)hp.com> --- src/bridge.c | 94 +++++++++++++++++++++------------------------------------- 1 files changed, 34 insertions(+), 60 deletions(-) diff --git a/src/bridge.c b/src/bridge.c index 0509afd..ec37192 100644 --- a/src/bridge.c +++ b/src/bridge.c @@ -49,8 +49,6 @@ #include "util.h" #include "logging.h" -#define MAX_TAP_ID 256 - #define JIFFIES_TO_MS(j) (((j)*1000)/HZ) #define MS_TO_JIFFIES(ms) (((ms)*HZ)/1000) @@ -466,76 +464,52 @@ brAddTap(brControl *ctl, int vnet_hdr, int *tapfd) { - int id, subst, fd; + int fd, len; + struct ifreq ifr = {0}; if (!ctl || !ctl->fd || !bridge || !ifname) return EINVAL; - subst = id = 0; - - if (strstr(*ifname, "%d")) - subst = 1; - if ((fd = open("/dev/net/tun", O_RDWR)) < 0) return errno; - if (vnet_hdr) - vnet_hdr = brProbeVnetHdr(fd); + ifr.ifr_flags = IFF_TAP|IFF_NO_PI; - do { - struct ifreq try; - int len; +#ifdef IFF_VNET_HDR + if (vnet_hdr && brProbeVnetHdr(fd)) + ifr.ifr_flags |= IFF_VNET_HDR; +#endif - memset(&try, 0, sizeof(struct ifreq)); + strncpy(ifr.ifr_name, *ifname, IFNAMSIZ-1); - try.ifr_flags = IFF_TAP|IFF_NO_PI; + if (ioctl(fd, TUNSETIFF, &ifr) < 0) + goto error; -#ifdef IFF_VNET_HDR - if (vnet_hdr) - try.ifr_flags |= IFF_VNET_HDR; -#endif + len = strlen(ifr.ifr_name); + if (len >= BR_IFNAME_MAXLEN - 1) { + errno = EINVAL; + goto error; + } - if (subst) { - len = snprintf(try.ifr_name, BR_IFNAME_MAXLEN, *ifname, id); - if (len >= BR_IFNAME_MAXLEN) { - errno = EADDRINUSE; - goto error; - } - } else { - len = strlen(*ifname); - if (len >= BR_IFNAME_MAXLEN - 1) { - errno = EINVAL; - goto error; - } - - strncpy(try.ifr_name, *ifname, len); - try.ifr_name[len] = '\0'; - } - - if (ioctl(fd, TUNSETIFF, &try) == 0) { - /* We need to set the interface MTU before adding it - * to the bridge, because the bridge will have its - * MTU adjusted automatically when we add the new interface. - */ - if ((errno = brSetInterfaceMtu(ctl, bridge, try.ifr_name))) - goto error; - if ((errno = brAddInterface(ctl, bridge, try.ifr_name))) - goto error; - if ((errno = brSetInterfaceUp(ctl, try.ifr_name, 1))) - goto error; - if (!tapfd && - (errno = ioctl(fd, TUNSETPERSIST, 1))) - goto error; - VIR_FREE(*ifname); - if (!(*ifname = strdup(try.ifr_name))) - goto error; - if (tapfd) - *tapfd = fd; - return 0; - } - - id++; - } while (subst && id <= MAX_TAP_ID); + /* We need to set the interface MTU before adding it + * to the bridge, because the bridge will have its + * MTU adjusted automatically when we add the new interface. + */ + if ((errno = brSetInterfaceMtu(ctl, bridge, ifr.ifr_name))) + goto error; + if ((errno = brAddInterface(ctl, bridge, ifr.ifr_name))) + goto error; + if ((errno = brSetInterfaceUp(ctl, ifr.ifr_name, 1))) + goto error; + if (!tapfd && + (errno = ioctl(fd, TUNSETPERSIST, 1))) + goto error; + VIR_FREE(*ifname); + if (!(*ifname = strdup(ifr.ifr_name))) + goto error; + if (tapfd) + *tapfd = fd; + return 0; error: close(fd);

3 3

[libvirt] VMware support and Server 2.0.1
by Timo Makinen 29 Jul '09

29 Jul '09

I tested libvirt with VMware Server 2.0.1 (not ESX). Server starts http on port 8333 as default and libvirt api did not have any option to give port number when connecting. Maybe connecting should be changed to support uris like: esx://host:123/ Next i tried to hardcode port into libvirt and got following error: error: internal error Expecting server major/minor version '3.5' or '4.0' but found '2.0.1' I patched libvirt again and finally got connected. Well at least list command gave out list of vms. With dumpxml i got following error: virsh # dumpxml vm02 error: internal error Missing essential config entry '' - Timo

2 1

[libvirt] [PATCH] qemu: fix monitor socket reconnection
by Ryota Ozaki 29 Jul '09

29 Jul '09

Hi, This patch fixes a handling of a qemu monitor unix socket. Current code addresses 'Socket may not have shown up yet' case with catching EACCES of connect(2), however, in that case, an errno which should be catched is ENOENT, not EACCES. Thanks, ozaki-r >From 5a4b7ee2296ca6756127c26c6bfbbe1f49ab8722 Mon Sep 17 00:00:00 2001 From: Ryota Ozaki <ozaki.ryota(a)gmail.com> Date: Tue, 28 Jul 2009 23:48:22 +0900 Subject: [PATCH] qemu: fix monitor socket reconnection --- src/qemu_driver.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/qemu_driver.c b/src/qemu_driver.c index 52a77be..9fb8506 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -917,8 +917,8 @@ qemudOpenMonitorUnix(virConnectPtr conn, if (ret == 0) break; - if (errno == EACCES || errno == ECONNREFUSED) { - /* EACCES : Socket may not have shown up yet + if (errno == ENOENT || errno == ECONNREFUSED) { + /* ENOENT : Socket may not have shown up yet * ECONNREFUSED : Leftover socket hasn't been removed yet */ continue; } -- 1.6.0.6

4 4

[libvirt] [PATCH] Kernel command line support for UML
by Ron Yorston 29 Jul '09

29 Jul '09

When experimenting with libvirt and UML I found it necessary to add support for additional kernel command line arguments. Ron diff -up libvirt-0.6.5/src/uml_conf.c.uml-cmdline libvirt-0.6.5/src/uml_conf.c --- libvirt-0.6.5/src/uml_conf.c.uml-cmdline 2009-07-10 16:28:51.000000000 +0100 +++ libvirt-0.6.5/src/uml_conf.c 2009-07-27 19:39:59.000000000 +0100 @@ -326,6 +326,31 @@ umlBuildCommandLineChr(virConnectPtr con } /* + * Null-terminate the current argument and return a pointer to the next. + * This should follow the same rules as the Linux kernel: arguments are + * separated by spaces; arguments can be quoted with double quotes; double + * quotes can't be escaped. + */ +static char *umlNextArg(char *args) +{ + int in_quote = 0; + + for (; *args; args++) { + if (*args == ' ' && !in_quote) { + *args++ = '\0'; + break; + } + if (*args == '"') + in_quote = !in_quote; + } + + while (*args == ' ') + args++; + + return args; +} + +/* * Constructs a argv suitable for launching uml with config defined * for a given virtual machine. */ @@ -342,6 +367,7 @@ int umlBuildCommandLine(virConnectPtr co const char **qargv = NULL; int qenvc = 0, qenva = 0; const char **qenv = NULL; + char *cmdline = NULL; uname(&ut); @@ -474,6 +500,22 @@ int umlBuildCommandLine(virConnectPtr co ADD_ARG(ret); } + if (vm->def->os.cmdline) { + char *args, *next_arg; + if ((cmdline = strdup(vm->def->os.cmdline)) == NULL) + goto no_memory; + + args = cmdline; + while (*args == ' ') + args++; + + while (*args) { + next_arg = umlNextArg(args); + ADD_ARG_LIT(args); + args = next_arg; + } + } + ADD_ARG(NULL); ADD_ENV(NULL); @@ -495,6 +537,7 @@ int umlBuildCommandLine(virConnectPtr co VIR_FREE((qenv)[i]); VIR_FREE(qenv); } + VIR_FREE(cmdline); return -1; #undef ADD_ARG

3 2

[libvirt] [PATCH 1/3] Move ldconfig calls to libvirt-client %post/%postun
by Mark McLoughlin 29 Jul '09

29 Jul '09

Now that the library has moved to libvirt-client --- libvirt.spec.in | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 6228d0b..b79e766 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -411,7 +411,6 @@ chmod 0644 $RPM_BUILD_ROOT%{_sysconfdir}/sysconfig/libvirtd rm -fr %{buildroot} %post -/sbin/ldconfig %if %{with_libvirtd} %if %{with_qemu} @@ -440,7 +439,9 @@ if [ $1 = 0 ]; then fi %endif -%postun -p /sbin/ldconfig +%post client -p /sbin/ldconfig + +%postun client -p /sbin/ldconfig %files %defattr(-, root, root) -- 1.6.2.5

2 5

[libvirt] [PATCH] libvirt-glib: add missing details argument to callbacks
by Guido Günther 29 Jul '09

29 Jul '09

Hi, the python example is lacking the details argument in the callback. This makes the example fail due to the wrong number of arguments. Cheers, -- Guido

2 5

[libvirt] Debug the username/passwords
by Bryan Kearney 29 Jul '09

29 Jul '09

Hopefully this is the last question I have on the Java drivers. I am trying to debug the callback from openAuth(). I am passing back what I think is valid information, but I see in /var/log/messages the following: Jul 28 16:21:24 localhost libvirtd: sasl step failed -20 (SASL(-13): user not found: no secret in database) Is there a way to output what I have passed back in the credential responses? Setting the following in the /etc/libvirt/libvirtd.conf file does not seem to increase the logging: log_level=1 log_outputs="1:stderr 1:syslog:bk 1:file:virsh.log" and setting LIBVIRT_DEBUG=1,and launching libvirtd from the command line does not spit out the info. -- bk

2 2

[libvirt] ANNOUNCE: New release virt-manager 0.8.0
by Cole Robinson 29 Jul '09

29 Jul '09

I'm happy to announce a new virt-manager release, version 0.8.0. The release can be downloaded from: http://virt-manager.org/download.html The direct download link is: http://virt-manager.org/download/sources/virt-manager/virt-manager-0.8.0.ta… This release includes: - New 'Clone VM' Wizard - Improved UI, including an overhaul of the main 'manager' view - System tray icon for easy VM access (start, stop, view console/details) - Wizard for adding serial, parallel, and video devices to existing VMs. - CPU pinning support (Michal Novotny) - Ability to view and change VM security (sVirt) settings (Dan Walsh) - Many bug fixes and improvements Thanks to everyone who has contributed to this release through testing, bug reporting, submitting patches, and otherwise sending in feedback! Thanks, Cole

4 3

[libvirt] PATCH: Fix remote domain events deadlock
by Daniel P. Berrange 29 Jul '09

29 Jul '09

There is a deadlock in the remote driver's dispatch code for domain events. If your callback triggers a API call that does an RPC call, then the driver will deadlock. The fix is the same as we did in the QEMU driver. We releae the lock & re-aquire it, with some extreme cleverness to protect against people deleting callbacks during the course of dispatch Daniel commit e2e6f795dd1832227c7d5c7832596be3642fe6e4 Author: Daniel P. Berrange <berrange(a)redhat.com> Date: Tue Jul 28 17:45:19 2009 +0100 Fix deadlock in remote driver domain events * src/remote_internal.c: Release driver lock when dispatching events to callbacks diff --git a/src/remote_internal.c b/src/remote_internal.c index f20ed6e..a58b768 100644 --- a/src/remote_internal.c +++ b/src/remote_internal.c @@ -167,6 +167,8 @@ struct private_data { virDomainEventQueuePtr domainEvents; /* Timer for flushing domainEvents queue */ int eventFlushTimer; + /* Flag if we're in process of dispatching */ + int domainEventDispatching; /* Self-pipe to wakeup threads waiting in poll() */ int wakeupSendFD; @@ -6288,18 +6290,26 @@ static int remoteDomainEventDeregister (virConnectPtr conn, remoteDriverLock(priv); - if (virDomainEventCallbackListRemove(conn, priv->callbackList, - callback) < 0) { - error (conn, VIR_ERR_RPC, _("removing cb fron list")); - goto done; - } - - if ( priv->callbackList->count == 0 ) { - /* Tell the server when we are the last callback deregistering */ - if (call (conn, priv, 0, REMOTE_PROC_DOMAIN_EVENTS_DEREGISTER, - (xdrproc_t) xdr_void, (char *) NULL, - (xdrproc_t) xdr_void, (char *) NULL) == -1) + if (priv->domainEventDispatching) { + if (virDomainEventCallbackListMarkDelete(conn, priv->callbackList, + callback) < 0) { + error (conn, VIR_ERR_RPC, _("marking cb for deletion")); goto done; + } + } else { + if (virDomainEventCallbackListRemove(conn, priv->callbackList, + callback) < 0) { + error (conn, VIR_ERR_RPC, _("removing cb from list")); + goto done; + } + + if ( priv->callbackList->count == 0 ) { + /* Tell the server when we are the last callback deregistering */ + if (call (conn, priv, 0, REMOTE_PROC_DOMAIN_EVENTS_DEREGISTER, + (xdrproc_t) xdr_void, (char *) NULL, + (xdrproc_t) xdr_void, (char *) NULL) == -1) + goto done; + } } rv = 0; @@ -7309,18 +7319,54 @@ done: remoteDriverUnlock(priv); } +static void remoteDomainEventDispatchFunc(virConnectPtr conn, + virDomainEventPtr event, + virConnectDomainEventCallback cb, + void *cbopaque, + void *opaque) +{ + struct private_data *priv = opaque; + + /* Drop the lock whle dispatching, for sake of re-entrancy */ + remoteDriverUnlock(priv); + virDomainEventDispatchDefaultFunc(conn, event, cb, cbopaque, NULL); + remoteDriverLock(priv); +} + void remoteDomainEventQueueFlush(int timer ATTRIBUTE_UNUSED, void *opaque) { virConnectPtr conn = opaque; struct private_data *priv = conn->privateData; + virDomainEventQueue tempQueue; remoteDriverLock(priv); - virDomainEventQueueDispatch(priv->domainEvents, priv->callbackList, - virDomainEventDispatchDefaultFunc, NULL); + priv->domainEventDispatching = 1; + + /* Copy the queue, so we're reentrant safe */ + tempQueue.count = priv->domainEvents->count; + tempQueue.events = priv->domainEvents->events; + priv->domainEvents->count = 0; + priv->domainEvents->events = NULL; + + virDomainEventQueueDispatch(&tempQueue, priv->callbackList, + remoteDomainEventDispatchFunc, priv); virEventUpdateTimeout(priv->eventFlushTimer, -1); + /* Purge any deleted callbacks */ + virDomainEventCallbackListPurgeMarked(priv->callbackList); + + if ( priv->callbackList->count == 0 ) { + /* Tell the server when we are the last callback deregistering */ + if (call (conn, priv, 0, REMOTE_PROC_DOMAIN_EVENTS_DEREGISTER, + (xdrproc_t) xdr_void, (char *) NULL, + (xdrproc_t) xdr_void, (char *) NULL) == -1) + VIR_WARN0("Failed to de-register events"); + } + + priv->domainEventDispatching = 0; + remoteDriverUnlock(priv); } -- |: 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 :|

2 1

[libvirt] [PATCH] Fix trailing whitespace in NEWS
by Mark McLoughlin 29 Jul '09

29 Jul '09

Hey, I had to commit this before I could push Cheers, Mark.

1 0

[libvirt] [PATCH 00/14] Clean up spec files mess
by Mark McLoughlin 29 Jul '09

29 Jul '09

Hey, I get treated like a war criminal when I make a change in the Fedora spec file and don't send it upstream, yet there are a whole heap of changes which haven't been pushed upstream ... only one of which is mine ... two changes less than DV himself :-) The following series of patches pulls in the changes from Fedora which make sense. Some of them I'm not 100% sure of: - defaulting to policy kit - is there some reason we don't want that upstream? - The python cruft thing, not sure its needed - Converting NEWS to UTF-8 could probably be done upstream - Requiring libselinux is dubious, doesn't autorequires handle that? My final contention is that maintaining these spec files in the upstream repo actually makes it harder to maintain the Fedora spec file and helps nobody. Because of half-assed attempts to keep the two in sync, I ended up having to fix a bunch of stuff and re-instate another bunch of stuff in Fedora: - Enable netcf support - Pass --with-qemu-user=qemu etc. to configure - Move various requires to the libvirt-client sub-package - Sync some trivial cleanups from upstream spec file - Remove explicit libxml2 requires, again - Build with --without-capng if capng support is disabled - Remove explicit dir creating in makeinstall, replaced by attr in files - Set perms on /var/{run,lib,cache}/libvirt/qemu I really think we should kill these off and just point people at the Fedora spec file if they're looking for a reference spec file. I need some very strong rationale before I ever go wasting my time (2 hours) on this again. Cheers, Mark.

4 23

[libvirt] [PATCH] Remove MAX_TAP_ID
by Aron Griffis 29 Jul '09

29 Jul '09

As far as I can tell, there's no reason to format the device string in brAddTap(). Delegate the job to TUNSETIFF, thereby removing the loop and the MAX_TAP_ID artificial limit. This patch allows me to get 421 guests running before hitting other limits. Signed-off-by: Aron Griffis <aron.griffis(a)hp.com> --- src/bridge.c | 94 +++++++++++++++++++++------------------------------------- 1 files changed, 34 insertions(+), 60 deletions(-) diff --git a/.gnulib b/.gnulib deleted file mode 160000 index b653eda..0000000 --- a/.gnulib +++ /dev/null @@ -1 +0,0 @@ -Subproject commit b653eda3ac4864de205419d9f41eec267cb89eeb diff --git a/src/bridge.c b/src/bridge.c index 0509afd..ec37192 100644 --- a/src/bridge.c +++ b/src/bridge.c @@ -49,8 +49,6 @@ #include "util.h" #include "logging.h" -#define MAX_TAP_ID 256 - #define JIFFIES_TO_MS(j) (((j)*1000)/HZ) #define MS_TO_JIFFIES(ms) (((ms)*HZ)/1000) @@ -466,76 +464,52 @@ brAddTap(brControl *ctl, int vnet_hdr, int *tapfd) { - int id, subst, fd; + int fd, len; + struct ifreq ifr = {0}; if (!ctl || !ctl->fd || !bridge || !ifname) return EINVAL; - subst = id = 0; - - if (strstr(*ifname, "%d")) - subst = 1; - if ((fd = open("/dev/net/tun", O_RDWR)) < 0) return errno; - if (vnet_hdr) - vnet_hdr = brProbeVnetHdr(fd); + ifr.ifr_flags = IFF_TAP|IFF_NO_PI; - do { - struct ifreq try; - int len; +#ifdef IFF_VNET_HDR + if (vnet_hdr && brProbeVnetHdr(fd)) + ifr.ifr_flags |= IFF_VNET_HDR; +#endif - memset(&try, 0, sizeof(struct ifreq)); + strncpy(ifr.ifr_name, *ifname, IFNAMSIZ-1); - try.ifr_flags = IFF_TAP|IFF_NO_PI; + if (ioctl(fd, TUNSETIFF, &ifr) < 0) + goto error; -#ifdef IFF_VNET_HDR - if (vnet_hdr) - try.ifr_flags |= IFF_VNET_HDR; -#endif + len = strlen(ifr.ifr_name); + if (len >= BR_IFNAME_MAXLEN - 1) { + errno = EINVAL; + goto error; + } - if (subst) { - len = snprintf(try.ifr_name, BR_IFNAME_MAXLEN, *ifname, id); - if (len >= BR_IFNAME_MAXLEN) { - errno = EADDRINUSE; - goto error; - } - } else { - len = strlen(*ifname); - if (len >= BR_IFNAME_MAXLEN - 1) { - errno = EINVAL; - goto error; - } - - strncpy(try.ifr_name, *ifname, len); - try.ifr_name[len] = '\0'; - } - - if (ioctl(fd, TUNSETIFF, &try) == 0) { - /* We need to set the interface MTU before adding it - * to the bridge, because the bridge will have its - * MTU adjusted automatically when we add the new interface. - */ - if ((errno = brSetInterfaceMtu(ctl, bridge, try.ifr_name))) - goto error; - if ((errno = brAddInterface(ctl, bridge, try.ifr_name))) - goto error; - if ((errno = brSetInterfaceUp(ctl, try.ifr_name, 1))) - goto error; - if (!tapfd && - (errno = ioctl(fd, TUNSETPERSIST, 1))) - goto error; - VIR_FREE(*ifname); - if (!(*ifname = strdup(try.ifr_name))) - goto error; - if (tapfd) - *tapfd = fd; - return 0; - } - - id++; - } while (subst && id <= MAX_TAP_ID); + /* We need to set the interface MTU before adding it + * to the bridge, because the bridge will have its + * MTU adjusted automatically when we add the new interface. + */ + if ((errno = brSetInterfaceMtu(ctl, bridge, ifr.ifr_name))) + goto error; + if ((errno = brAddInterface(ctl, bridge, ifr.ifr_name))) + goto error; + if ((errno = brSetInterfaceUp(ctl, ifr.ifr_name, 1))) + goto error; + if (!tapfd && + (errno = ioctl(fd, TUNSETPERSIST, 1))) + goto error; + VIR_FREE(*ifname); + if (!(*ifname = strdup(ifr.ifr_name))) + goto error; + if (tapfd) + *tapfd = fd; + return 0; error: close(fd);

1 0

[libvirt] ANNOUNCE: New release virtinst 0.500.0
by Cole Robinson 29 Jul '09

29 Jul '09

I'm happy to announce a new virtinst release, version 0.500.0. The release can be downloaded from: http://virt-manager.org/download.html The direct download link is: http://virt-manager.org/download/sources/virtinst/virtinst-0.500.0.tar.gz This release includes: - New virt-install device options --serial, --parallel, and --video - Allow various auth types for libvirt connections (PolicyKit, SASL, ...) - New virt-clone option --auto-clone: generates all needed input. - Option to specify network device model via virt-install --network (Guido Gunther) - New virt-install option --virt-type for specifying hypervisor type (kvm, qemu). --accelerate is now the default behavior: To provision a plain qemu VM on a KVM enabled host, use '--virt-type qemu' - OVF input support for virt-convert - Many bug fixes and improvements Thanks to everyone who has contributed to this release through testing, bug reporting, submitting patches, and otherwise sending in feedback! Thanks, Cole

1 0

[libvirt] [PATCH] qemu: fix missing ifname in qemu argument
by Ryota Ozaki 28 Jul '09

28 Jul '09

Hi, This patch adds missing ifname, which is a tap name for network/bridge configuration, in qemu argument of -net. Without this patch, a specified tap name, e.g., <target dev='tap01'/>, in a domain XML will be ignored. Thanks, ozaki-r >From feeca4f2cecf5c6836f7aec6e004585f3b855bed Mon Sep 17 00:00:00 2001 From: Ryota Ozaki <ozaki.ryota(a)gmail.com> Date: Tue, 28 Jul 2009 23:46:12 +0900 Subject: [PATCH] qemu: fix missing ifname in qemu argument --- src/qemu_conf.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/qemu_conf.c b/src/qemu_conf.c index 4fe4e39..48bf928 100644 --- a/src/qemu_conf.c +++ b/src/qemu_conf.c @@ -1187,11 +1187,13 @@ qemuBuildHostNetStr(virConnectPtr conn, switch (net->type) { case VIR_DOMAIN_NET_TYPE_NETWORK: case VIR_DOMAIN_NET_TYPE_BRIDGE: - if (virAsprintf(str, "%stap%cfd=%s,vlan=%d%s%s", + if (virAsprintf(str, "%stap%cfd=%s,vlan=%d%s%s%s%s", prefix ? prefix : "", type_sep, tapfd, vlan, (net->hostnet_name ? ",name=" : ""), - (net->hostnet_name ? net->hostnet_name : "")) < 0) { + (net->hostnet_name ? net->hostnet_name : ""), + (net->ifname ? ",ifname=" : ""), + (net->ifname ? net->ifname : "")) < 0) { virReportOOMError(conn); return -1; } -- 1.6.0.6

2 2

[libvirt] Java JNA Patch series
by Bryan Kearney 28 Jul '09

28 Jul '09

Per the earlier email, here is the patch series

2 32

Re: [libvirt] VMware support and libcurl on rhel-u1
by Shahar Klein 28 Jul '09

28 Jul '09

apperantly I needed to unset CURLOPT_SSL_VERIFYHOST as well patched that and it is working now 10x Shahar ________________________________ From: Tom Hughes <tom(a)compton.nu> To: Shahar Klein <shaharklein(a)yahoo.com> Sent: Monday, July 27, 2009 2:36:32 PM Subject: Re: [libvirt] VMware support and libcurl on rhel-u1 On 27/07/09 12:33, Shahar Klein wrote: > this is what I thought but it doesn't work for me: > > [root@rain8 libvirt]# virsh -c esx://rain3?transport=http > Enter username for rain3 [root]: > Enter root password for rain3: > error: internal error curl_easy_perform() returned an error: SSL peer > certificate was not ok (51) > error: failed to connect to the hypervisor > > maybe because I'm using curl 7.15? That looks like it is still checking the certificate. Are you sure you're not still running the unpatched one somehow? I'm running curl 7.19.4 which isn't that different - does the curl_easy_setopt manual page mention the CURLOPT_SSL_VERIFYPEER option? Tom -- Tom Hughes (tom(a)compton.nu) http://www.compton.nu/

4 5

[libvirt] [PATCH 2/2] RPM spec file updated with glusterfs dependency
by Harshavardhana 28 Jul '09

28 Jul '09

Add new dependency for glusterfs rpm. --- libvirt.spec.in | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 1e8a8ef..ccef3a5 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -86,6 +86,8 @@ BuildRequires: util-linux # For showmount in FS driver (netfs discovery) BuildRequires: nfs-utils Requires: nfs-utils +# For glusterfs +Requires: glusterfs-client >= 2.0.2 %endif %if %{with_qemu} # From QEMU RPMs -- 1.6.0.6

4 6

[libvirt] [RFC/Experimental]: Tunnelled migration
by Chris Lalancette 28 Jul '09

28 Jul '09

All, Attached is the current version of the tunnelled migration patch, based upon danpb's generic datastream work. In order to use this work, you must first grab danpb's data-streams git branch here: http://gitorious.org/~berrange/libvirt/staging and then apply this patch on top. In some basic testing, this seems to work fine for me, although I have not given it a difficult scenario nor measured CPU utilization with these patches in place. DanB, these patches take a slightly different approach than you and I discussed yesterday on IRC. Just to recap, you suggested a new version of virMigratePrepare (called virMigratePrepareTunnel) that would take in as one of the arguments a datastream, and during the prepare step properly setup the datastream. Unless I'm missing something (which is entirely possible), this would also require passing that same datastream into the perform and finish stages, meaning that I'd essentially have an all new migration protocol version 3. To try to avoid that, during the prepare I store the port that we used to start the listening qemu in a new field in the virDomainObj structure. Then during the perform step, I create a datastream on the destination and run a new RPC function called virDomainMigratePrepareTunnel. This looks that port back up, associates it with the current stream, and returns back to the caller. Then the source side just does virStreamSend for all the data, and we have tunnelled migration. TODO: - More testing, especially under worst-case scenarios (VM constantly changing it's memory during migration) - CPU utilization testing to make sure that we aren't using a lot of CPU time doing this - Wall-clock testing - Switch over to using Unix Domain Sockets instead of localhost TCP migration. With a patch I put into upstream qemu (and is now in F-12), we can totally get rid of the scanning of localhost ports to find a free one, and just use Unix Domain Sockets. That should make the whole thing more robust. -- Chris Lalancette

2 2

[libvirt] test driver and virtOpenAuth()
by Bryan Kearney 28 Jul '09

28 Jul '09

I am trying to test a call into virConnectOpenAuth. How can I configure the test driver to accept auth calls on "test+tcp://localhost/default" Thanks! -- bk

2 4

[libvirt] [RFC] secrets API, was Re: [PATCH 0/9] Add support for (qcow*) volume encryption
by Miloslav Trmac 28 Jul '09

28 Jul '09

Hello, based on your comments, here is a proposal for a "secret management" API. Rather than add explicit accessors for attributes of secrets, and hard-code the "secrets are related to storage volumes" association in the API, the proposed uses XML to manipulate the association as well as other attributes, like it is used in other areas of libvirt. The user would allocate an ID for the secret using virSecretAllocateID(), set attributes of the secret using XML, e.g. <secret ephemeral='no' private='yes'> <volume>/var/lib/libvirt/images/mail.img</volume> <description>LUKS passphrase for the main hard drive of our mail server</description> </secret> Then, the secret value can be either generated and stored by libvirt, or supplied by the user using virSecretSetValue(). A simple API is provided for enumeration of all secrets. Very large deployments manage secret IDs automatically, so it probably is not necessary to provide specialized lookup functions (e.g. allowing the volume key -> secret ID lookup in less than O(number of secrets)). The <encryption> element used in volume and domain specifications remains, but it never contains any secrets directly, only something like <secret type='passphrase' secret_id='c1f11a6d-8c5d-4a3e-ac7a-4e171c5e0d4a' /> More detailed documentation is in the patch. Does that look OK? Thank you, Mirek

1 0

Re: [libvirt] [PATCH 0/9] Add support for (qcow*) volume encryption
by Miloslav Trmac 28 Jul '09

28 Jul '09

Hello, ----- "Daniel P. Berrange" <berrange(a)redhat.com> wrote: > On Fri, Jul 24, 2009 at 07:25:54AM -0400, Miloslav Trmac wrote: > > A client in this case is the central, fully trusted, management > > system (e.g. oVirt), there is no need to protect against it. > > A more likely flow is > > > > MGMT client (no knowledge of secrets) > > | > > v > > MGMT server + key server (integrated or separate but cooperating) > > | > > v > > libvirt daemon > > | > > v > > qemu > > > > > What I am suggesting is that libvirt daemon should communicate > > > with the key server directly in all cases, and take the client > > > out of the loop. The client should merely indicate whether it > > > wants encryption or not, and never be able to directly access > > > any key material itself. With a direct trust relationship > > > between the key server and each libvirtd daemon, you do now > > > have a guarentee that keys are only ever used on the node for > > > which they are intended. You also have the additional guarentee > > > that no libvirt client can ever see any key secrets or passphrases > > > since it has been taken completely out of the loop. > > > > As far as I understand it, the whole point of virtual machine > > encryption is that the nodes are _not_ trusted, and different > > encryption keys protect data on different nodes. > > I did not mean to imply that libvirtd on a node should have > access to *all* secrets. Each libvirtd daemon has its own > identity, and when talking to a keys server it would authenticate > and the key server would only allow it access to some sub-set > of keys. ie, only the keys neccessary for VMs it needs to run. <snip> > If you have each libvirtd requesting secrets directly from the > keystore, at the time it starts a guest, then should an admin > issue a migrate command manually, the destination libvirtd > would still be unable to access the secrets of the incoming > VM, since the keystore will not have been configured to allow > it access. > > We also have to bear in mind how the MGMT server communicates > with the libvirt daemon. One likely option is using a messaging > service, which offers authenticity but not secrecy. ie, libvirt > receiving a request off the message bus can be sure the request > came from the mgmtm server, but cannot be sure it wasn't seen > by other users during transport. Thus by including secrets in > the XML, you could be exposing them to the message bus adminstrator. Is this likely to happen? No such transport is currently implemented AFAICS, and just using TLS is much simpler - it is already implemented, and gives you authentication of both sides almost for free. Why would such a messaging service be necessary for the client<->libvirtd connection, but not the libvirtd<->key server connection? > Taking your diagram, I think I would generalize it still further > to allow for mgmt server to be optional separate from key server, > and to introduce a "message bus" between MGMT server & libvirt. > Really pushing the limits of ASCII art... > > MGMT client > | > V > MGMT server <---> Key server > | ^ > V | > message bus | > | | > V | > libvirt daemon <----/ > | > V > QEMU This makes sense if you do not want to store any secrets locally at the node at any cost, but the strong coupling of the libvirtd daemon and the key server is a significant disadvantage: not only would the key server and libvirtd both have to implement the same protocol for transferring secrets, but the MGMT server, key server and libvirtd would all have to share the same concept of "identity" and the same method of authentication - and the MGMT server would have to manipulate the libvirt accounts on the key server. There are a few products that offer a company-wide key server, but there AFAIK is no standardized protocol for transferring the secrets yet (the KMIP committee was formed only a few months ago), and not even a proposal for any account management on the key server. We don't know what the industry standard will look like - or whether there will be any; right now, any key server used in connection with libvirt-managed nodes would have to implement a libvirt-specified interface (in addition to any interface it may provide to other clients). It is much simpler to use the already implemented libvirt remote interface, and give the MGMT server implementors a free hand in deciding which key server, if any, they will use, and to avoid the non-trivial effort of specifying the specialized libvirtd<->key server protocol and implementing both ends of it. If secrets related to auto-start domains are stored locally on the node, all secrets needed by libvirtd will be anticipated by the MGMT server, who can provide them before starting the operation. This makes the libvirtd->key server connection unnecessary. The API proposal that will follow adds a secret management interface to libvirtd. The API allows libvirtd to use a persistent local secret store, as well as an external key server as described above. In addition, it also allows a "fully managed" mode, where the storage and provision of secrets is the responsibility of the MGMT server. This is accomplished by giving each secret two attributes: "private", which prohibits libvirtd from divulging the secret to any client (this was done in the previous implementation in XML, but now that secrets are out-of-bound, making secrets private does not break any round-trip XML editing), and "ephemeral", which prohibits libvirt from storing the secret persistently. The MGMT server would send the necessary secrets before any operation that needs them, and delete them from the node when they are no longer necessary. Looking at your use cases: > - A node can only see secrets for VMs it is configured to run Same in "fully managed" mode. > - MGMT server does not need to ever see the secrets itself. It > merely controls access for which nodes can see them, and can > request generation of new secrets The MGMT server needs to see the secrets in "fully managed mode". The MGMT server can actually see them even if libvirtd talks to the key server directly, because the key server is allowed to manipulate permissions of the secrets and it can give itself the permission to read them. The MGMT server can also delete any interface, or boot a "rescue image" that copies the plaintext out of the virtual machine. The MGMT server must be trusted in any case. > - Messages between the MGMT server & libvirtd do not need to > be encrypted, since they don't include secrets I can't see why this is useful - the connection will use TLS anyway. > - Other users who authenticate to libvirt on a node cannot > move a VM to an unauthorized node, since it can't see secrets Same in "fully managed" mode. > - VM save memory images (which include the fully XML) do not ever > expose the secrets. So VM save image cannot be restored on an > unauthorized node. Same in "fully managed" mode. Mirek

1 0

[libvirt] VMware ESX: Add no_verify query parameter
by Matthias Bolte 27 Jul '09

27 Jul '09

Adds a no_verify query parameter to stop libcurl from verifying the server certificate for the https transport. Idea from Tom Hughes. Regards, Matthias

3 2

[libvirt] PATCH: Implement vCPU cpuTime + cpu fields for QEMU
by Daniel P. Berrange 27 Jul '09

27 Jul '09

The qemudDomainGetVcpus() method in QEMU driver has the following long standing todo item /* XXX cpu time, current pCPU mapping */ This has caused confusion for users, because they set affinity and then wonder why 'virsh vcpuinfo' constantly reports their guest as running on pCPU 0. This patch implements the missing bits, pulling it out of /proc/$PID/task/$TID/stat ie, the per-vCPU thread status file Daniel diff --git a/src/qemu_driver.c b/src/qemu_driver.c index c594495..388843f 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -2465,24 +2465,44 @@ cleanup: } -static int qemudGetProcessInfo(unsigned long long *cpuTime, int pid) { +static int qemudGetProcessInfo(unsigned long long *cpuTime, int *lastCpu, int pid, int tid) { char proc[PATH_MAX]; FILE *pidinfo; unsigned long long usertime, systime; + int cpu; + int ret; - if (snprintf(proc, sizeof(proc), "/proc/%d/stat", pid) >= (int)sizeof(proc)) { + if (tid) + ret = snprintf(proc, sizeof(proc), "/proc/%d/task/%d/stat", pid, tid); + else + ret = snprintf(proc, sizeof(proc), "/proc/%d/stat", pid); + if (ret >= (int)sizeof(proc)) { + errno = E2BIG; return -1; } if (!(pidinfo = fopen(proc, "r"))) { /*printf("cannot read pid info");*/ /* VM probably shut down, so fake 0 */ - *cpuTime = 0; + if (cpuTime) + *cpuTime = 0; + if (lastCpu) + *lastCpu = 0; return 0; } - if (fscanf(pidinfo, "%*d %*s %*c %*d %*d %*d %*d %*d %*u %*u %*u %*u %*u %llu %llu", &usertime, &systime) != 2) { - qemudDebug("not enough arg"); + /* See 'man proc' for information about what all these fields are. We're + * only interested in a very few of them */ + if (fscanf(pidinfo, + /* pid -> stime */ + "%*d %*s %*c %*d %*d %*d %*d %*d %*u %*u %*u %*u %*u %llu %llu" + /* cutime -> endcode */ + "%*d %*d %*d %*d %*d %*u %*u %*d %*u %*u %*u %*u" + /* startstack -> processor */ + "%*u %*u %*u %*u %*u %*u %*u %*u %*u %*u %*d %d", + &usertime, &systime, &cpu) != 3) { + VIR_WARN0("cannot parse process status data"); + errno = -EINVAL; return -1; } @@ -2491,9 +2511,14 @@ static int qemudGetProcessInfo(unsigned long long *cpuTime, int pid) { * _SC_CLK_TCK is jiffies per second * So calulate thus.... */ - *cpuTime = 1000ull * 1000ull * 1000ull * (usertime + systime) / (unsigned long long)sysconf(_SC_CLK_TCK); + if (cpuTime) + *cpuTime = 1000ull * 1000ull * 1000ull * (usertime + systime) / (unsigned long long)sysconf(_SC_CLK_TCK); + if (lastCpu) + *lastCpu = cpu; + - qemudDebug("Got %llu %llu %llu", usertime, systime, *cpuTime); + VIR_DEBUG("Got status for %d/%d user=%llu sys=%llu cpu=%d", + pid, tid, usertime, systime, cpu); fclose(pidinfo); @@ -3133,7 +3158,7 @@ static int qemudDomainGetInfo(virDomainPtr dom, if (!virDomainIsActive(vm)) { info->cpuTime = 0; } else { - if (qemudGetProcessInfo(&(info->cpuTime), vm->pid) < 0) { + if (qemudGetProcessInfo(&(info->cpuTime), NULL, vm->pid, 0) < 0) { qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, ("cannot read cputime for domain")); goto cleanup; } @@ -3676,7 +3701,16 @@ qemudDomainGetVcpus(virDomainPtr dom, for (i = 0 ; i < maxinfo ; i++) { info[i].number = i; info[i].state = VIR_VCPU_RUNNING; - /* XXX cpu time, current pCPU mapping */ + + if (vm->vcpupids != NULL && + qemudGetProcessInfo(&(info[i].cpuTime), + &(info[i].cpu), + vm->pid, + vm->vcpupids[i]) < 0) { + virReportSystemError(dom->conn, errno, "%s", + _("cannot get vCPU placement & pCPU time")); + goto cleanup; + } } } -- |: 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 :|

2 1

[libvirt] [PATCH] Canonicalize qemu machine types
by Mark McLoughlin 27 Jul '09

27 Jul '09

In qemu-0.11 there is a 'pc-0.10' machine type which allows you to run guests with a machine which is compatible with the pc machine in qemu-0.10 - e.g. using the original PCI class for virtio-blk and virtio-console and disabling MSI support in virtio-net. The idea here is that we don't want to surprise guests by changing the hardware when qemu is updated. I've just posted some patches for qemu-0.11 which allows libvirt to canonicalize the 'pc' machine alias to the latest machine version. This patches makes us use that so that when a guest is configured to use the 'pc' machine type, we resolve that to 'pc-0.11' machine and save that in the guest XML. See also: https://fedoraproject.org/wiki/Features/KVM_Stable_Guest_ABI * src/qemu_conf.c: add qemudCanonicalizeMachine() to parse the output of 'qemu -M ?' * src/qemu_driver.c: canonicalize the machine type in qemudDomainDefine() --- src/qemu_conf.c | 114 +++++++++++++++++++++++++++++++++++++++++++++++++++++ src/qemu_conf.h | 3 + src/qemu_driver.c | 3 + 3 files changed, 120 insertions(+), 0 deletions(-) diff --git a/src/qemu_conf.c b/src/qemu_conf.c index 4043d70..3f4edfa 100644 --- a/src/qemu_conf.c +++ b/src/qemu_conf.c @@ -470,6 +470,120 @@ virCapsPtr qemudCapsInit(void) { return NULL; } +/* Format is: + * <machine> <desc> [(alias of <machine>)] + */ +static int +qemudParseMachineTypesStr(const char *machines ATTRIBUTE_UNUSED, + const char *machine ATTRIBUTE_UNUSED, + char **canonical) +{ + const char *p = machines; + + *canonical = NULL; + + do { + const char *eol; + char *s; + + if (!(eol = strchr(p, '\n'))) + return -1; /* eof file without finding @machine */ + + if (!STRPREFIX(p, machine)) { + p = eol + 1; + continue; /* doesn't match @machine */ + } + + p += strlen(machine); + + if (*p != ' ') { + p = eol + 1; + continue; /* not a complete match of @machine */ + } + + do { + p++; + } while (*p == ' '); + + p = strstr(p, "(alias of "); + if (!p || p > eol) + return 0; /* not an alias, name is canonical */ + + *canonical = strndup(p + strlen("(alias of "), eol - p); + + s = strchr(*canonical, ')'); + if (!s) { + VIR_FREE(*canonical); + *canonical = NULL; + return -1; /* output is screwed up */ + } + + *s = '\0'; + break; + } while (1); + + return 0; +} + +int +qemudCanonicalizeMachine(virConnectPtr conn ATTRIBUTE_UNUSED, + virDomainDefPtr def) +{ + const char *const qemuarg[] = { def->emulator, "-M", "?", NULL }; + const char *const qemuenv[] = { "LC_ALL=C", NULL }; + char *machines, *canonical; + enum { MAX_MACHINES_OUTPUT_SIZE = 1024*4 }; + pid_t child; + int newstdout = -1, len; + int ret = -1, status; + + if (virExec(NULL, qemuarg, qemuenv, NULL, + &child, -1, &newstdout, NULL, VIR_EXEC_CLEAR_CAPS) < 0) + return -1; + + len = virFileReadLimFD(newstdout, MAX_MACHINES_OUTPUT_SIZE, &machines); + if (len < 0) { + virReportSystemError(NULL, errno, "%s", + _("Unable to read 'qemu -M ?' output")); + goto cleanup; + } + + if (qemudParseMachineTypesStr(machines, def->os.machine, &canonical) < 0) + goto cleanup2; + + if (canonical) { + VIR_FREE(def->os.machine); + def->os.machine = canonical; + } + + ret = 0; + +cleanup2: + VIR_FREE(machines); +cleanup: + if (close(newstdout) < 0) + ret = -1; + +rewait: + if (waitpid(child, &status, 0) != child) { + if (errno == EINTR) + goto rewait; + + VIR_ERROR(_("Unexpected exit status from qemu %d pid %lu"), + WEXITSTATUS(status), (unsigned long)child); + ret = -1; + } + /* Check & log unexpected exit status, but don't fail, + * as there's really no need to throw an error if we did + * actually read a valid version number above */ + if (WEXITSTATUS(status) != 0) { + VIR_WARN(_("Unexpected exit status '%d', qemu probably failed"), + WEXITSTATUS(status)); + } + + return ret; +} + static unsigned int qemudComputeCmdFlags(const char *help, unsigned int version, unsigned int is_kvm, diff --git a/src/qemu_conf.h b/src/qemu_conf.h index fbf2ab9..b668669 100644 --- a/src/qemu_conf.h +++ b/src/qemu_conf.h @@ -123,6 +123,9 @@ int qemudExtractVersionInfo (const char *qemu, unsigned int *version, unsigned int *flags); +int qemudCanonicalizeMachine (virConnectPtr conn, + virDomainDefPtr def); + int qemudParseHelpStr (const char *str, unsigned int *flags, unsigned int *version, diff --git a/src/qemu_driver.c b/src/qemu_driver.c index d2db1a2..98f8e95 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -4102,6 +4102,9 @@ static virDomainPtr qemudDomainDefine(virConnectPtr conn, const char *xml) { } } + if (qemudCanonicalizeMachine(conn, def) < 0) + goto cleanup; + if (!(vm = virDomainAssignDef(conn, &driver->domains, def))) { -- 1.6.2.5

4 27

[libvirt] Unable to get libvirt working with KVM using direct kernel boot.
by Tiit Kaeeli 27 Jul '09

27 Jul '09

Hi, I am trying to get libvirt to manage a KVM virtual machine on Debian squeeze with multiple relevant parts included form unstable (because of need for gpt partition tables, Intel 82576 nic etc). Unfortunately the VM does not seem to boot. I will see a quemu prompt on /dev/pts/1. (Tried pressing c on it, but no difference). Nothing will appear on /dev/pts/2, /dev/pts/3 or when I type virsh console vm1_storage. If I grab the kvm command line generated by libvirt using ps and run it manually, it works fine. (Starts booting after I press c on /dev/pts/1 or remove the -S option from the command line) Libvirt configuration file: <domain type='kvm'> <name>vm1_storage</name> <uuid>29bfac01-b24d-e4ab-e741-f33f7e880d9d</uuid> <memory>4096000</memory> <currentMemory>4096000</currentMemory> <vcpu>6</vcpu> <os> <type>hvm</type> <kernel>/boot/vmlinuz-2.6.30-1-amd64</kernel> <initrd>/boot/initrd.img-2.6.30-1-amd64</initrd> <cmdline>"root=UUID=98d6d3d7-3782-4f6b-a94f-bc0272c0289d ro console=tty0 console=ttyS0,115200n8"</cmdline> </os> <features> <acpi/> <pae/> </features> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>restart</on_crash> <devices> <emulator>/usr/bin/kvm</emulator> <disk type='block' device='disk'> <source dev='/dev/vm_lvm/vm01_storage_root'/> <target dev='vda' bus='virtio'/> </disk> <interface type='ethernet'> <mac address='52:54:00:12:34:56'/> <target dev='tap0'/> </interface> <serial type='pty'> <target port='0'/> </serial> <console type='pty'> <target port='0'/> </console> </devices> </domain> KVM command line that I see in ps after running virsh define /etc/libvirt/qemu/vm1_storage.xml virsh start vm1_storage (Works fine after I remove the -S option or press 'c' on quemu monitor on /dev/pts/1): /usr/bin/kvm -S -M pc -m 4000 -smp 6 -name vm1_storage -uuid 29bfac01-b24d-e4ab-e741-f33f7e880d9d -nographic -monitor pty -boot c -kernel /boot/vmlinuz-2.6.30-1-amd64 -initrd /boot/initrd.img-2.6.30-1-amd64 -append "root=UUID=98d6d3d7-3782-4f6b-a94f-bc0272c0289d ro console=tty0 console=ttyS0,115200n8" -drive file=/dev/vm_lvm/vm01_storage_root,if=virtio,index=0,boot=on -net nic,macaddr=52:54:00:12:34:56,vlan=0 -net tap,ifname=tap0,vlan=0 -serial pty -parallel none -usb /var/log/libvirt/qemu/vm1_storage.log shows: LC_ALL=C PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin HOME=/ /usr/bin/kvm -S -M pc -m 4000 -smp 6 -name vm1_storage -uuid 29bfac01-b24d-e4ab-e741-f33f7e880d9d -nographic -monitor pty -boot c -kernel /boot/vmlinuz-2.6.30-1-amd64 -initrd /boot/initrd.img-2.6.30-1-amd64 -append "root=UUID=98d6d3d7-3782-4f6b-a94f-bc0272c0289d ro console=tty0 console=ttyS0,115200n8" -drive file=/dev/vm_lvm/vm01_storage_root,if=virtio,index=0,boot=on -net nic,macaddr=52:54:00:12:34:56,vlan=0 -net tap,ifname=tap0,vlan=0 -serial pty -parallel none -usb char device redirected to /dev/pts/1 char device redirected to /dev/pts/2 qemu: loading initrd (0x76055b bytes) at 0x000000007f89f000 Versions in use: kvm 85+dfsg-4 libvirt0 0.6.5-2 libvirt-bin 0.6.5-2 linux-image-2.6.30-1-amd64 Thanks for any ideas.

2 1

[libvirt] VMware ESX: Allow leading dots in VMX config entry names
by Matthias Bolte 27 Jul '09

27 Jul '09

Hi, '.encoding = "UTF-8"' is a valid VMX config entry, so the virConfParser must accept it in VMX mode. This patch lets the parser accept those names with leading dots. Regards, Matthias

3 5

[libvirt] VMware ESX: Accept VI API version 4.0
by Matthias Bolte 27 Jul '09

27 Jul '09

Hi, This patch extends the VI API version checks to accept version 4.0 too. Regards, Matthias

2 2

[libvirt] VMware ESX: Fix memory leaks in esxDomainDumpXML()
by Matthias Bolte 27 Jul '09

27 Jul '09

Subject says it all.

2 1

[libvirt] Unable to get libvirt working with KVM using direct kernel boot.
by Tiit Kaeeli 27 Jul '09

27 Jul '09

Hi, I am trying to get libvirt to manage a KVM virtual machine on Debian squeeze with multiple relevant parts included form unstable (because of need for gpt partition tables, Intel 82576 nic etc). Unfortunately the VM does not seem to boot. I will see a quemu prompt on /dev/pts/1. (Tried pressing c on it, but no difference). Nothing will appear on /dev/pts/2, /dev/pts/3 or when I type virsh console vm1_storage. If I grab the kvm command line generated by libvirt using ps and run it manually, it works fine. (Starts booting after I press c on /dev/pts/1 or remove the -S option from the command line) Libvirt configuration file: <domain type='kvm'> <name>vm1_storage</name> <uuid>29bfac01-b24d-e4ab-e741-f33f7e880d9d</uuid> <memory>4096000</memory> <currentMemory>4096000</currentMemory> <vcpu>6</vcpu> <os> <type>hvm</type> <kernel>/boot/vmlinuz-2.6.30-1-amd64</kernel> <initrd>/boot/initrd.img-2.6.30-1-amd64</initrd> <cmdline>"root=UUID=98d6d3d7-3782-4f6b-a94f-bc0272c0289d ro console=tty0 console=ttyS0,115200n8"</cmdline> </os> <features> <acpi/> <pae/> </features> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>restart</on_crash> <devices> <emulator>/usr/bin/kvm</emulator> <disk type='block' device='disk'> <source dev='/dev/vm_lvm/vm01_storage_root'/> <target dev='vda' bus='virtio'/> </disk> <interface type='ethernet'> <mac address='52:54:00:12:34:56'/> <target dev='tap0'/> </interface> <serial type='pty'> <target port='0'/> </serial> <console type='pty'> <target port='0'/> </console> </devices> </domain> KVM command line that I see in ps after running virsh define /etc/libvirt/qemu/vm1_storage.xml virsh start vm1_storage (Works fine after I remove the -S option or press 'c' on quemu monitor on /dev/pts/1): /usr/bin/kvm -S -M pc -m 4000 -smp 6 -name vm1_storage -uuid 29bfac01-b24d-e4ab-e741-f33f7e880d9d -nographic -monitor pty -boot c -kernel /boot/vmlinuz-2.6.30-1-amd64 -initrd /boot/initrd.img-2.6.30-1-amd64 -append "root=UUID=98d6d3d7-3782-4f6b-a94f-bc0272c0289d ro console=tty0 console=ttyS0,115200n8" -drive file=/dev/vm_lvm/vm01_storage_root,if=virtio,index=0,boot=on -net nic,macaddr=52:54:00:12:34:56,vlan=0 -net tap,ifname=tap0,vlan=0 -serial pty -parallel none -usb /var/log/libvirt/qemu/vm1_storage.log shows: LC_ALL=C PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin HOME=/ /usr/bin/kvm -S -M pc -m 4000 -smp 6 -name vm1_storage -uuid 29bfac01-b24d-e4ab-e741-f33f7e880d9d -nographic -monitor pty -boot c -kernel /boot/vmlinuz-2.6.30-1-amd64 -initrd /boot/initrd.img-2.6.30-1-amd64 -append "root=UUID=98d6d3d7-3782-4f6b-a94f-bc0272c0289d ro console=tty0 console=ttyS0,115200n8" -drive file=/dev/vm_lvm/vm01_storage_root,if=virtio,index=0,boot=on -net nic,macaddr=52:54:00:12:34:56,vlan=0 -net tap,ifname=tap0,vlan=0 -serial pty -parallel none -usb char device redirected to /dev/pts/1 char device redirected to /dev/pts/2 qemu: loading initrd (0x76055b bytes) at 0x000000007f89f000 Versions in use: kvm 85+dfsg-4 libvirt0 0.6.5-2 libvirt-bin 0.6.5-2 linux-image-2.6.30-1-amd64 Thanks for any ideas.

1 0

Re: [libvirt] VMware support and libcurl on rhel-u1
by Shahar Klein 27 Jul '09

27 Jul '09

I'll try with 7.15 can I test with ESX4i? Shahar ________________________________ From: Matthias Bolte <matthias.bolte(a)googlemail.com> To: Shahar Klein <shaharklein(a)yahoo.com> Sent: Sunday, July 26, 2009 5:56:18 PM Subject: Re: [libvirt] VMware support and libcurl on rhel-u1 2009/7/26 Shahar Klein <shaharklein(a)yahoo.com>: > It doesn't because I built libcurl on a different path (/usr/local/lib) > So I won't damage other stuff on this box. > I thought that this is what the LIBCURL_LIBS and LIBCURL_CFLAGS env vars are > for. No, but installing in the /usr/local prefix should also work with pkg-config. The .pc for libcurl should be located at /usr/local/lib/pkgconfig/libcurl.pc and pkg-config should find it. Maybe you have to set PKG_CONFIG_PATH in order to have pkg-config pick it up. Check the output of: pkg-config libcurl --modversion It should output 7.18 or what ever your version is, if installed correctly. > btw - Do I really need libcurl 7.18? trying to install it pushing me into > the depnendncy hell: [...] Hm... I just looked at the changelist of libcurl and something older like version 7.15 may also work, but I haven't tested it. You could give it a try and edit the configure.in file and change 7.18 to 7.15 and see if that works out. Regards, Matthias

3 10

[libvirt] [PATCH] Power Hypervisor support
by Eduardo Otubo 27 Jul '09

27 Jul '09

Hello everyone, This should be the official patch for the libvrt-0.7.0 release. Here I'll comment all the features already implemented and the roadmap we have ahead: Features supported: * Connects to HMC/VIOS or IVM systems. * Life cycle commands (resume and shutdown). * dumpxml * 'list' and 'list --all' What is being implemented: * better and centralized control for UUID * definexml * CPU management commands Any comment are always welcome. []'s -- Eduardo Otubo Software Engineer Linux Technology Center IBM Systems & Technology Group Mobile: +55 19 8135 0885 otubo(a)linux.vnet.ibm.com

3 5

[libvirt] [PATCH] test: Implement BlockStats and InterfaceStats
by Cole Robinson 27 Jul '09

27 Jul '09

We fake stats values based on the current time, similar to how it's done for cpu time. --- src/test.c | 116 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 files changed, 114 insertions(+), 2 deletions(-) diff --git a/src/test.c b/src/test.c index f02b58b..6db3825 100644 --- a/src/test.c +++ b/src/test.c @@ -2093,6 +2093,118 @@ cleanup: return ret; } +static int testDomainBlockStats(virDomainPtr domain, + const char *path, + struct _virDomainBlockStats *stats) +{ + testConnPtr privconn = domain->conn->privateData; + virDomainObjPtr privdom; + struct timeval tv; + unsigned long long statbase; + int i, found = 0, ret = -1; + + testDriverLock(privconn); + privdom = virDomainFindByName(&privconn->domains, + domain->name); + testDriverUnlock(privconn); + + if (privdom == NULL) { + testError(domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__); + goto error; + } + + for (i = 0 ; i < privdom->def->ndisks ; i++) { + if (STREQ(path, privdom->def->disks[i]->dst)) { + found = 1; + break; + } + } + + if (!found) { + testError(domain->conn, VIR_ERR_INVALID_ARG, + _("invalid path: %s"), path); + goto error; + } + + if (gettimeofday(&tv, NULL) < 0) { + virReportSystemError(domain->conn, errno, + "%s", _("getting time of day")); + goto error; + } + + /* No significance to these numbers, just enough to mix it up*/ + statbase = (tv.tv_sec * 1000UL * 1000UL) + tv.tv_usec; + stats->rd_req = statbase / 10; + stats->rd_bytes = statbase / 20; + stats->wr_req = statbase / 30; + stats->wr_bytes = statbase / 40; + stats->errs = tv.tv_sec / 2; + + ret = 0; +error: + if (privdom) + virDomainObjUnlock(privdom); + return ret; +} + +static int testDomainInterfaceStats(virDomainPtr domain, + const char *path, + struct _virDomainInterfaceStats *stats) +{ + testConnPtr privconn = domain->conn->privateData; + virDomainObjPtr privdom; + struct timeval tv; + unsigned long long statbase; + int i, found = 0, ret = -1; + + testDriverLock(privconn); + privdom = virDomainFindByName(&privconn->domains, + domain->name); + testDriverUnlock(privconn); + + if (privdom == NULL) { + testError(domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__); + goto error; + } + + for (i = 0 ; i < privdom->def->nnets ; i++) { + if (privdom->def->nets[i]->ifname && + STREQ (privdom->def->nets[i]->ifname, path)) { + ret = 0; + break; + } + } + + if (!found) { + testError(domain->conn, VIR_ERR_INVALID_ARG, + _("invalid path, '%s' is not a known interface"), path); + goto error; + } + + if (gettimeofday(&tv, NULL) < 0) { + virReportSystemError(domain->conn, errno, + "%s", _("getting time of day")); + goto error; + } + + /* No significance to these numbers, just enough to mix it up*/ + statbase = (tv.tv_sec * 1000UL * 1000UL) + tv.tv_usec; + stats->rx_bytes = statbase / 10; + stats->rx_packets = statbase / 100; + stats->rx_errs = tv.tv_sec / 1; + stats->rx_drop = tv.tv_sec / 2; + stats->tx_bytes = statbase / 20; + stats->tx_packets = statbase / 110; + stats->tx_errs = tv.tv_sec / 3; + stats->tx_drop = tv.tv_sec / 4; + + ret = 0; +error: + if (privdom) + virDomainObjUnlock(privdom); + return ret; +} + static virDrvOpenStatus testOpenNetwork(virConnectPtr conn, virConnectAuthPtr auth ATTRIBUTE_UNUSED, int flags ATTRIBUTE_UNUSED) { @@ -4117,8 +4229,8 @@ static virDriver testDriver = { NULL, /* domainMigratePrepare */ NULL, /* domainMigratePerform */ NULL, /* domainMigrateFinish */ - NULL, /* domainBlockStats */ - NULL, /* domainInterfaceStats */ + testDomainBlockStats, /* domainBlockStats */ + testDomainInterfaceStats, /* domainInterfaceStats */ NULL, /* domainBlockPeek */ NULL, /* domainMemoryPeek */ testNodeGetCellsFreeMemory, /* nodeGetCellsFreeMemory */ -- 1.6.0.6

2 3

[libvirt] VMware support and libcurl on rhel-u1
by Shahar Klein 26 Jul '09

26 Jul '09

I'm trying to play with the new ESX driver on a rhel5-u1 x86_64 but I get this: checking for libcurl libraries >= 7.18.0... configure: error: libcurl >= 7.18.0 is required for the ESX driver according to http://curl.haxx.se/download.html there is no rpm for rhel x86_64 so I built it and tried: export LIBCURL_LIBS="-L/usr/local/lib -lcurl -L/usr/kerberos/lib64" export LIBCURL_CFLAGS="-I/usr/local/include" but configure didn't care about it am I doing something wrong here? tnx Shahar Klein

2 1

[libvirt] How to get access to QEMU monitor?
by Jun Koi 26 Jul '09

26 Jul '09

Hi, I am running QEMU via libvirt (using virsh & virt-manager). Now I want to get access to the monitor interface of QEMU, so I can issue some commands to monitor. Is there anyway to do that? Thanks, J

2 1

[libvirt] VMware ESX driver progress
by Matthias Bolte 25 Jul '09

25 Jul '09

Hi, the development was hindered by our testing cluster being offline since 2 weeks due to server room cooling system maintenance. But I finally got a basic version of dump-XML done, that fills all fields of the virDomainDef that the VMX file contains data for. Changes since first announcement: - Move code into esx subdirectory - Add esxNodeGetInfo() - Fix esxDomainGetInfo() to report the correct value for memory - Add memory and max-memory getter/setters - Add CPU scheduler getter/setters - Validate a migration before trying to perform it - Replace esxUtil_Strdup() with strdup() and remove esxUtil_MigrateStringFromLibXML() - Add esxVI_EnsureSession() to handle expiring sessions - Separate VI client code into multiple files and generate most of the type handling code with macros - Add esxDomainDumpXML() based on esxVMX_ParseConfig() The ESX driver isn't complete yet, currently it supports: - domain lookup by ID, UUID and name - domain listing - domain info retrieval - domain suspend and resume - domain start and destroy - domain reboot and shutdown, if the VMware tools are installed inside the domain - domain migration with previous validation - domain memory configuration - domain CPU amount and scheduler configuration - domain XML-dump - domain XML-from-native (VMX) - node info retrieval - node hostname retrieval The next milestone is creating/defining a new domain from a domain XML config, but this will not be finished until feature freeze for 0.7.0 on friday. Regards, Matthias

3 9

[libvirt] [PATCH 1/1] Support for Callbacks and VirtualBox Version 3.
by Pritesh Kothari 24 Jul '09

24 Jul '09

Hi All, I have added support for callbacks and VirtualBox version 3. I have attached the patch for the same, it is made against the head as of today. To check if the callbacks work you can run the example from examples/domain- events/events-c directory and then start a domain to see if you get any callback events. Regards, Pritesh

3 4

Re: [libvirt] [PATCH 0/9] Add support for (qcow*) volume encryption
by Miloslav Trmac 24 Jul '09

24 Jul '09

----- "Daniel P. Berrange" <berrange(a)redhat.com> wrote: > > Not quite: the main case of a "dumb" client would be a large-scale > > virtualization management software that contains a primary store > > of encryption information, and gives each node access only to those > > keys that are currently necessary by the node to run its domains; > > the fact that each node has access to only a limited set of keys > > prevents an attacker that compromises a single node from reading > > disk images of all domains managed in the entire site, even if > > the disk image storage (e.g. unencrypted NFS) does not allow > > managing access by each node separately. > > > > Such a client must be able to transfer the actual secrets, not only > > identifiers, to libvirt. (The idea of a "dumb" client, that does > > not know the specifics of the format, is an additional feature on > > top, but one that implies that the client does send the actual > secrets.) > > This implies a flow of secrets > > Key server --\ > +-> libvirt client -> libvirt daemon -> qemu > MGMT server --/ > > This does not in fact guarentee that secrets for a particular > node are only used on the node for which they are intended, > because the key server cannot be sure of what libvirt daemons > the libvirt client is connected to. A client in this case is the central, fully trusted, management system (e.g. oVirt), there is no need to protect against it. A more likely flow is MGMT client (no knowledge of secrets) | v MGMT server + key server (integrated or separate but cooperating) | v libvirt daemon | v qemu > What I am suggesting is that libvirt daemon should communicate > with the key server directly in all cases, and take the client > out of the loop. The client should merely indicate whether it > wants encryption or not, and never be able to directly access > any key material itself. With a direct trust relationship > between the key server and each libvirtd daemon, you do now > have a guarentee that keys are only ever used on the node for > which they are intended. You also have the additional guarentee > that no libvirt client can ever see any key secrets or passphrases > since it has been taken completely out of the loop. As far as I understand it, the whole point of virtual machine encryption is that the nodes are _not_ trusted, and different encryption keys protect data on different nodes. If all nodes are trusted, what additional functionality does volume encryption with per-volume keys provide? If the nodes are trusted to read only data from the domains they currently host, the nodes could just as well use an encrypted local hard drive to store all images, or share a single key to encrypt all images stored on a NFS/SAN. > Key server > | > V > MGMT server -> libvirt client -> libvirt daemon -> qemu > > > Storage of secrets in a separate keystore is more important for "local" > > libvirt deployments, where libvirt manages the primary, long-term, > > store of the secrets. > > Nope, I think that having a keystore for all scenarios is the > desirable end goal. Passing secrets around in the XML should > ideally never be done at all - we should aim to always have a > keystore that can be used. Secrets in the XML would just be a > fallback for a rarely used niche cases, or disaster recovery, > or experimentation. That means that any deployment of more than one node with migration requires a separate server providing a shared keystore, even if there is only one client to manage the nodes. The N nodes x N "management consoles" case requires a centralized key store, but it's not necessary to impose it on the "1 management console" case. > > > 2. A desktop key agent (eg gnome-keyring) > > > > > > This would be useful for the unprivileged libvirtd instances > > > that run in the context of the desktop login session. Users > > > already have SSH, GPG, website keys stored here, so having > > > keys for their VM disks is obviously desirable > > > > Another option here is to let the client store the secrets in > > gnome-keyring, and transfer them to libvirt only when starting > > a domain (especially when there are no persistent domains). > > That doesn't affect the design in any way, though. > > This is undesriable because it lets requires that any client which > wishes to start the guest must have access to the secrets. We really > need to be able to have separation of this, so that when we introduce > fine grained access control, you can setup separate roles for users > who can access / work with secrets, vs users who can start/define > guests. > > > > 3. A off-node key management server > > > > > > This would be useful for a large scale data center / cluster > > > cloud deployment. This is good to allow management > scalability > > > and better separation of responsiblities of adminstration. > > > > > > If no keystore is in use, then I clearly all keys must go in & > out > > > of libvirt using the XML, which is pretty much what you're doing > > > in this series. I would say though, there is no point in clearing > > > the secret from the virStorageVolDefPtr instance/XML after volume > > > creation, since the secret is going to be kept in memory forever > > > for any guest using the volume. By not clearing the secret, an > > > app could create a volume, requesting automatic key generation, > > > then just do virStorageVolDumpXML(vol, > VIR_STORAGE_VOL_XML_SECURE) > > > to extract it and pass to onto the XML for the guest that's > created > > > > That is unreliable with current implementation, because a pool refresh > > creates all information about volumes anew by reading the volume files; > > therefore, after any pool refresh libvirt "forgets" the secrets directly > > associated with a volume (not secrets associated with an use of a volume > > in a domain). If client A creates a volume, client B can refresh the > > pool before client A is able to read the automatically generated secrets. > > One of the reasons the patch is clearing the information immediately is > > to make sure a {virStorageVolCreateXML; virStorageVolGetXMLDesc) always > > fails and no client is written depending on this racy operation sequence. > > > There is another, more important, reason why the node should never return > > encryption data to anyone who can connect to it. Consider the above- > > described situation of a large-scale virtualization deployment that uses > > volume image encryption to restrict access of nodes to volume data: If > > domain migration is supported, the nodes (all of them, or at least nodes > > within some groups) must be able to connect "read-write" to other nodes. > > This is just another argument for taking clients out of the loop completely > and having libvirt daemon always talk directly with keystores. > > > With a keystore we'd likely need a handful of APIs > > > > > > - create a secret providing a passphrase > > > - list all known secret UUIDs > > > - get the passphrase assoicated with a secret > > > - delete a secret based on UUID > > > - lookup a secret UUID based on the a disk path > > > > We might not need all of these. There are three main use > > cases for storing keys outside of XML: > > > > 1) This is a "small-scale deployment", where libvirt is the > > primary key store, and anyone that is able to connect to > > libvirtd is implicitly allowed to use the secrets. In this > > case the keystore can be managed completely automatically - > > creating a volume, or perhaps using it for the first time, > > implies storing a secret, and deleting a volume implies > > deleting a secret. Secrets are identified using a volume- > > unique object (is that a path, or an UUID?), which is > > completely transparent to the client. As long as the volume > > is "known" to libvirt, the client does not even need to > > specify any <encryption> element when creating a domain. > > > > (1a, a "small-scale deployment" where there are multiple > > clients that should be protected against each other, and > > libvirt is the primary key store, does not make sense > > unless a client account system is added to libvirt. > > > 2) The secrets are primarily stored outside libvirt (to > > restrict node access to image data), and libvirt stores > > only secrets for currently defined persistent domains (to > > support domain autostart). In this case the keystore can > > be managed completely automatically - persistent domain > > definition implies storing a secret, deleting the domain > > implies deleting a secret. Secrets are identified using > > a (domain, volume) pair, and this identifier is never > > exposed to the client. > > Deleting the domain does not imply deleting a secret, since > secrets are really associated with disks, not domains. It does, in this case - if the domain is not running on the node, the secret should not be stored on the node at all. > A disk may be shared by multiple domains. If a disk is shared, there will be multiple (domain volume) key ID pairs, and deleting one instace of the secret does not delete the other. > In addition, when you delete > a domain, libvirt does not delete the disk. It is the clients > responsiblity to delete disks after the fact. In this scenario the client does the long-term key storage. libvirt does not need to - and shouldn't - store the key merely because a volume exists. > In any case items 1 & 2 are really just 2 different implementations > of the same concept. There is a keystore, and libvirt talks to > it. Whether the keystore is local, or remote from the node is a > minor impl detail. Whether libvirt can request any secret, even if does not need to know it, is not a minor detail. > Use of a keystore of some format should be our primary goal here, > since it takes the client out of the loop. When libvirt does ACLs > on clients this is even more important, because when you revoke a > clients' access to libvirt you can be sure they don't have a > record of any of the secrets, since they never had any opportunity > to see / use them during the time they were authenticated > > 3) An external key store (such as a KMIP server) is used. In this > > case the management of access rights, listing and deleting secrets, > > would be performed by interacting with the external key store > > directly. Secrets are identified by using the external key store > > identifiers, and the client and libvirtd send these identifiers to > > each other. > > As long as we allow for secrets to be passed in the XML it will always > be possible for a client to talk to a keystore, obtain the secrets and > pass them in the XML. This is really not a desirable model to aim for > by default though, because it requires clients to know about all the > different types of keystore. If a client does not know about a particular > type of keystore, then it becomes unable to manage guests on that > libvirtd. > Not to mention the issue of trust of the client & revocation of > access. > Yes, libvirt will grow the concept of a user in the future, as well as > access controls on (user, object, operation) tuples Overall, it seems this all boils down to one thing - is libvirt intended to be a simple "virtualization server" that locally performs requested operations, managed by a separate "virtualization management" client that has full knowledge about the site (nodes, storage, secrets, access rights), with clients connecting to the "virtualization management" software instead of libvirtd, or a complete "virtualization management solution" that integrates all site-wide knowedge, accessed by dumb clients using the libvirt protocol? My reading of http://www.libvirt.org/goals.html implies the former; you seem to talk about the latter model. Have I misunderstood the role of libvirt in a large-scale deployment? Thank you, Mirek

2 2

[libvirt] [PATCH] test: Generate net interface names when assigning XML.
by Cole Robinson 24 Jul '09

24 Jul '09

We need interface names to implement InterfaceStats. --- src/test.c | 76 +++++++++++++++++++++++++++++++++++++++++++++++++++++------ 1 files changed, 68 insertions(+), 8 deletions(-) diff --git a/src/test.c b/src/test.c index 098d235..f02b58b 100644 --- a/src/test.c +++ b/src/test.c @@ -228,6 +228,66 @@ static const unsigned long long defaultPoolAlloc = 0; static int testStoragePoolObjSetDefaults(virConnectPtr conn, virStoragePoolObjPtr pool); +static char * +testDomainGenerateIfname(virConnectPtr conn, + virDomainDefPtr domdef) { + int maxif = 1024; + int ifctr, i; + + for (ifctr = 0; ifctr < maxif; ++ifctr) { + char *ifname; + int found = 0; + + if (virAsprintf(&ifname, "testnet%d", ifctr) < 0) { + virReportOOMError(conn); + return NULL; + } + + /* Generate network interface names */ + for (i = 0 ; i < domdef->nnets ; i++) { + if (domdef->nets[i]->ifname && + STREQ (domdef->nets[i]->ifname, ifname)) { + found = 1; + break; + } + } + + if (!found) + return ifname; + } + + testError(conn, VIR_ERR_INTERNAL_ERROR, + _("Exceeded max iface limit %d"), maxif); + return NULL; +} + +static virDomainObjPtr +testDomainAssignDef(virConnectPtr conn, + virDomainObjList *domlist, + virDomainDefPtr domdef) +{ + virDomainObjPtr domobj = NULL; + int i = 0; + + for (i = 0; i < domdef->nnets; i++) { + char *ifname; + if (domdef->nets[i]->ifname) + continue; + + ifname = testDomainGenerateIfname(conn, domdef); + if (!ifname) + goto error; + + domdef->nets[i]->ifname = ifname; + } + + if (!(domobj = virDomainAssignDef(conn, domlist, domdef))) + goto error; + +error: + return domobj; +} + static int testOpenDefault(virConnectPtr conn) { int u; struct timeval tv; @@ -282,7 +342,7 @@ static int testOpenDefault(virConnectPtr conn) { defaultDomainXML, VIR_DOMAIN_XML_INACTIVE))) goto error; - if (!(domobj = virDomainAssignDef(conn, &privconn->domains, domdef))) { + if (!(domobj = testDomainAssignDef(conn, &privconn->domains, domdef))) { virDomainDefFree(domdef); goto error; } @@ -608,7 +668,7 @@ static int testOpenFromFile(virConnectPtr conn, goto error; } - if (!(dom = virDomainAssignDef(conn, &privconn->domains, def))) { + if (!(dom = testDomainAssignDef(conn, &privconn->domains, def))) { virDomainDefFree(def); goto error; } @@ -920,8 +980,8 @@ testDomainCreateXML(virConnectPtr conn, const char *xml, VIR_DOMAIN_XML_INACTIVE)) == NULL) goto cleanup; - if ((dom = virDomainAssignDef(conn, &privconn->domains, - def)) == NULL) { + if ((dom = testDomainAssignDef(conn, &privconn->domains, + def)) == NULL) { virDomainDefFree(def); goto cleanup; } @@ -1471,8 +1531,8 @@ static int testDomainRestore(virConnectPtr conn, if (!def) goto cleanup; - if ((dom = virDomainAssignDef(conn, &privconn->domains, - def)) == NULL) + if ((dom = testDomainAssignDef(conn, &privconn->domains, + def)) == NULL) goto cleanup; dom->state = VIR_DOMAIN_RUNNING; @@ -1763,8 +1823,8 @@ static virDomainPtr testDomainDefineXML(virConnectPtr conn, VIR_DOMAIN_XML_INACTIVE)) == NULL) goto cleanup; - if ((dom = virDomainAssignDef(conn, &privconn->domains, - def)) == NULL) { + if ((dom = testDomainAssignDef(conn, &privconn->domains, + def)) == NULL) { goto cleanup; } def = NULL; -- 1.6.0.6

2 2

[libvirt] [PATCH] Don't allow NULL paths for BlockStats and InterfaceStats
by Cole Robinson 24 Jul '09

24 Jul '09

Do the check in libvirt.c, to save drivers from the burden. This changes behavior slightly in the qemu driver: we no longer explictly error if passed an empty string. An error will still be thrown when the device lookup fails. --- src/libvirt.c | 4 ++-- src/qemu_driver.c | 6 ------ 2 files changed, 2 insertions(+), 8 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index 0e6d88f..d086052 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -3578,7 +3578,7 @@ virDomainBlockStats (virDomainPtr dom, const char *path, virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__); return -1; } - if (!stats || size > sizeof stats2) { + if (!path || !stats || size > sizeof stats2) { virLibDomainError (dom, VIR_ERR_INVALID_ARG, __FUNCTION__); goto error; } @@ -3636,7 +3636,7 @@ virDomainInterfaceStats (virDomainPtr dom, const char *path, virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__); return -1; } - if (!stats || size > sizeof stats2) { + if (!path || !stats || size > sizeof stats2) { virLibDomainError (dom, VIR_ERR_INVALID_ARG, __FUNCTION__); goto error; } diff --git a/src/qemu_driver.c b/src/qemu_driver.c index c4683ae..c8f451d 100644 --- a/src/qemu_driver.c +++ b/src/qemu_driver.c @@ -5329,12 +5329,6 @@ qemudDomainInterfaceStats (virDomainPtr dom, goto cleanup; } - if (!path || path[0] == '\0') { - qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG, - "%s", _("NULL or empty path")); - goto cleanup; - } - /* Check the path is one of the domain's network interfaces. */ for (i = 0 ; i < vm->def->nnets ; i++) { if (vm->def->nets[i]->ifname && -- 1.6.0.6

2 1

[libvirt] [PATCH] python: Raise exceptions if virDomain*Stats fail.
by Cole Robinson 24 Jul '09

24 Jul '09

The generator couldn't tell that the stats return values were pointers. Stick a white list in the function which tries to make this distinction. --- python/generator.py | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/python/generator.py b/python/generator.py index 9528839..feff7a3 100755 --- a/python/generator.py +++ b/python/generator.py @@ -683,7 +683,10 @@ functions_list_exception_test = { functions_list_default_test = "%s is None" def is_list_type (name): - return name[-1:] == "*" + whitelist = [ "virDomainBlockStats", + "virDomainInterfaceStats" ] + + return name[-1:] == "*" or name in whitelist def nameFixup(name, classe, type, file): # avoid a desastrous clash -- 1.6.0.6

2 1

[libvirt] [PATCH] document telnet serial console options
by Guido Günther 24 Jul '09

24 Jul '09

Hi, attached patch updates the console docs a bit. O.k. to apply? Cheers, -- Guido

2 1

Re: [libvirt] [PATCH 1/9] Add volume encryption information handling.
by Miloslav Trmac 24 Jul '09

24 Jul '09

----- "Daniel P. Berrange" <berrange(a)redhat.com> wrote: > For the domain XML I agree that libirt would not need to worry about > multiple LUKS keys, but for the storage volume XML we would need to > expose the fact that there are multiple keys,or allow creation of > volumes with multiple keys. I don't know. Does a fact that a feature exist imply that libvirt must fully expose it? Any use case for multiple LUKS passphrases I can think of is a bit far-fetched. > > We know that the type and amount of information that will need to be > > stored will vary depending on the "encryption format"; on the other > > hand, expecting that the information consists of independent "secrets", > > each with a simple and format-independent definition, is probably too > > optimistic. (As mentioned above, we might need a key slot ID associated > > with a LUKS passphrase; we might also need a password hash algorithm > > associated with a dm-crypt passphrase. The encryption formats used in > > practice are often complex enough that a "simple passphrase" can not > > be used.) Once we have to assume that each secret can have an "encryption > > format"-dependent format, I think it is much clearer to use something like > > The idea of a 'key slot' and 'password hash algorithm' are not neccessarily > unique to LUKS though. No, but the specific semantics and value ranges are very likely to be different across formats. > If we get 2 different encryption formats both requiring > the concept of a 'key slot' we need to represent them in the same way in the > XML, not hve a different XML for each format. I was arguing about making the internal representation format-specific, not the XML. In the XML, having something like <secret type='...' id='...'> and <parameter name='...' id='...'> is quite reasonable. I think this is not the case in the internal representation - the information will eventually have to be "parsed" into format-specific variables, and it seems better to me to do this at once from the XML, than to create an additional internal representation and additionl "parsing" code. <snip> > The libvirt approach to XML representation is to try and define XML format > that are indenpedant of specific implementations. This does imply that the > XML parser cannot really do anything beyond basic syntactic validation, > near zero semantic validation. The task of semantic validation of the > parsed data, is thus passed off to the internal drivers which provide the > specific implementations. Not having any driver-specific knowledge in the generic parser does make sense; I'll change the code. Mirek

1 0