June 2011 - Devel - libvirt List Archives

Re: [libvirt] [Qemu-devel] [PATCH 6/9] net: Improve layout of 'info network'
by Anthony Liguori 07 Jun '11

07 Jun '11

On 06/07/2011 11:45 AM, Jan Kiszka wrote: > Improve the layout when listing non-vlan clients via 'info network'. The > result looks like this: > > (qemu) info network > Devices not on any VLAN: > orphan: net=10.0.2.0, restricted=n > virtio-net-pci.0: model=virtio-net-pci,macaddr=52:54:00:12:34:56 > \ network2: fd=5 > e1000.0: model=e1000,macaddr=52:54:00:12:34:57 > \ network1: net=10.0.2.0, restricted=n > rtl8139.0: model=rtl8139,macaddr=52:54:00:12:34:58 > > ie. peers are grouped, orphans are listed as before. > > CC: Markus Armbruster<armbru(a)redhat.com> > Signed-off-by: Jan Kiszka<jan.kiszka(a)siemens.com> There isn't a query-network yet in QMP so libvirt is probably still using the HMP version. Can someone on the libvirt side Ack/Nack about whether this patch will break libvirt? Regards, Anthony Liguori > --- > net.c | 14 +++++++++----- > 1 files changed, 9 insertions(+), 5 deletions(-) > > diff --git a/net.c b/net.c > index 4f777c3..606ce70 100644 > --- a/net.c > +++ b/net.c > @@ -1224,7 +1224,8 @@ int do_netdev_del(Monitor *mon, const QDict *qdict, QObject **ret_data) > void do_info_network(Monitor *mon) > { > VLANState *vlan; > - VLANClientState *vc; > + VLANClientState *vc, *peer; > + net_client_type type; > > QTAILQ_FOREACH(vlan,&vlans, next) { > monitor_printf(mon, "VLAN %d devices:\n", vlan->id); > @@ -1235,11 +1236,14 @@ void do_info_network(Monitor *mon) > } > monitor_printf(mon, "Devices not on any VLAN:\n"); > QTAILQ_FOREACH(vc,&non_vlan_clients, next) { > - monitor_printf(mon, " %s: %s", vc->name, vc->info_str); > - if (vc->peer) { > - monitor_printf(mon, " peer=%s", vc->peer->name); > + peer = vc->peer; > + type = vc->info->type; > + if (!peer || type == NET_CLIENT_TYPE_NIC) { > + monitor_printf(mon, " %s: %s\n", vc->name, vc->info_str); > + } > + if (peer&& type == NET_CLIENT_TYPE_NIC) { > + monitor_printf(mon, " \\ %s: %s\n", peer->name, peer->info_str); > } > - monitor_printf(mon, "\n"); > } > } >

2 1

[libvirt] [PATCH 0/7] lxc: Improve container startup error reporting
by Cole Robinson 07 Jun '11

07 Jun '11

Currently most LXC container startup failures result in a reported success followed immediately by a guest shutoff. This patch series tries to improve the situation by adding a basic handshake from driver<->controller and controller<->container. If the handshake fails, the driver will return the output from the guest logfile, similar to how we do for qemu guests. A few additional error reporting improvements are also included. Cole Robinson (7): lxc: Drop container stdio as late as possible lxc: Don't report error in Wait/SendContinue lxc: Refactor controller command building lxc: Improve guest startup error reporting lxc: controller: Improve container error reporting lxc: Verify root fs exists before mounting lxc: Ensure container <init> actually exists src/lxc/lxc_container.c | 67 +++++++++++------- src/lxc/lxc_container.h | 2 + src/lxc/lxc_controller.c | 63 ++++++++++++++++- src/lxc/lxc_driver.c | 173 +++++++++++++++++++++++++++++++++++++--------- 4 files changed, 243 insertions(+), 62 deletions(-) -- 1.7.4.4

2 15

[libvirt] [PATCH 0/4 v3] More virCommand conversions and cleanups
by Cole Robinson 07 Jun '11

07 Jun '11

The following series converts the last user of virExec to use virCommand. The remaining functionality is then moved out of util.c and into command.c v2: Committed some patches Dropped final patch, can be submitted seperately Addressed Eric's comments v3: Commited some patches Addressed Eric's v2 comments Cole Robinson (4): qemu: Convert virExec usage to virCommand util: Remove unused virExec wrapper util: Implement virRun as a wrapper around virCommand Move virRun, virExec*, virFork to util/command cfg.mk | 3 +- src/libvirt_private.syms | 6 +- src/lxc/veth.c | 2 +- src/nwfilter/nwfilter_ebiptables_driver.c | 1 + src/qemu/qemu_driver.c | 26 +- src/qemu/qemu_process.c | 2 +- src/storage/storage_backend_fs.c | 2 +- src/storage/storage_backend_logical.c | 2 +- src/util/command.c | 565 ++++++++++++++++++++++++- src/util/command.h | 14 + src/util/ebtables.c | 2 +- src/util/pci.c | 2 +- src/util/util.c | 661 +---------------------------- src/util/util.h | 32 -- src/vmware/vmware_driver.c | 1 + 15 files changed, 608 insertions(+), 713 deletions(-) -- 1.7.4.4

2 10

[libvirt] [PATCH] Fix allocation of veth's to not skip an index
by Daniel P. Berrange 07 Jun '11

07 Jun '11

The algorithm for autoassigning vethXXX devices, was always skipping over the starting dev index when finding a free name for the guest device. This should only be done if the host device was autoallocated. * src/lxc/veth.c: Don't skip over veth indexes --- src/lxc/veth.c | 14 +++++++++----- 1 files changed, 9 insertions(+), 5 deletions(-) diff --git a/src/lxc/veth.c b/src/lxc/veth.c index a00aa23..22dbed5 100644 --- a/src/lxc/veth.c +++ b/src/lxc/veth.c @@ -46,6 +46,7 @@ static int getFreeVethName(char **veth, int startDev) int devNum = startDev-1; char *path = NULL; + VIR_DEBUG("Find free from veth%d", startDev); do { VIR_FREE(path); ++devNum; @@ -53,6 +54,7 @@ static int getFreeVethName(char **veth, int startDev) virReportOOMError(); return -1; } + VIR_DEBUG("Probe %s", path); } while (virFileExists(path)); VIR_FREE(path); @@ -60,6 +62,7 @@ static int getFreeVethName(char **veth, int startDev) virReportOOMError(); return -1; } + return devNum; } @@ -98,18 +101,19 @@ int vethCreate(char** veth1, char** veth2) bool veth1_alloc = false; bool veth2_alloc = false; - VIR_DEBUG("veth1: %s veth2: %s", NULLSTR(*veth1), NULLSTR(*veth2)); + VIR_DEBUG("Host: %s guest: %s", NULLSTR(*veth1), NULLSTR(*veth2)); if (*veth1 == NULL) { if ((vethDev = getFreeVethName(veth1, vethDev)) < 0) goto cleanup; - VIR_DEBUG("Assigned veth1: %s", *veth1); + VIR_DEBUG("Assigned host: %s", *veth1); veth1_alloc = true; + vethDev++; } argv[3] = *veth1; while (*veth2 == NULL) { - if ((vethDev = getFreeVethName(veth2, vethDev + 1)) < 0) { + if ((vethDev = getFreeVethName(veth2, vethDev)) < 0) { if (veth1_alloc) VIR_FREE(*veth1); goto cleanup; @@ -122,12 +126,12 @@ int vethCreate(char** veth1, char** veth2) continue; } - VIR_DEBUG("Assigned veth2: %s", *veth2); + VIR_DEBUG("Assigned guest: %s", *veth2); veth2_alloc = true; } argv[8] = *veth2; - VIR_DEBUG("veth1: %s veth2: %s", *veth1, *veth2); + VIR_DEBUG("Create Host: %s guest: %s", *veth1, *veth2); if (virRun(argv, NULL) < 0) { if (veth1_alloc) VIR_FREE(*veth1); -- 1.7.4.4

2 1

[libvirt] [PATCH] esx: Remove duplicated invalid-argument checks
by Matthias Bolte 07 Jun '11

07 Jun '11

Those checks are already performed at the public API level. --- src/esx/esx_driver.c | 10 ---------- src/esx/esx_storage_driver.c | 5 ----- 2 files changed, 0 insertions(+), 15 deletions(-) diff --git a/src/esx/esx_driver.c b/src/esx/esx_driver.c index 500dc52..a5b96a9 100644 --- a/src/esx/esx_driver.c +++ b/src/esx/esx_driver.c @@ -1479,11 +1479,6 @@ esxListDomains(virConnectPtr conn, int *ids, int maxids) esxVI_VirtualMachinePowerState powerState; int count = 0; - if (ids == NULL || maxids < 0) { - ESX_ERROR(VIR_ERR_INVALID_ARG, "%s", _("Invalid argument")); - return -1; - } - if (maxids == 0) { return 0; } @@ -2895,11 +2890,6 @@ esxListDefinedDomains(virConnectPtr conn, char **const names, int maxnames) int count = 0; int i; - if (names == NULL || maxnames < 0) { - ESX_ERROR(VIR_ERR_INVALID_ARG, "%s", _("Invalid argument")); - return -1; - } - if (maxnames == 0) { return 0; } diff --git a/src/esx/esx_storage_driver.c b/src/esx/esx_storage_driver.c index 6461917..c7fc992 100644 --- a/src/esx/esx_storage_driver.c +++ b/src/esx/esx_storage_driver.c @@ -167,11 +167,6 @@ esxListStoragePools(virConnectPtr conn, char **const names, int maxnames) int count = 0; int i; - if (names == NULL || maxnames < 0) { - ESX_ERROR(VIR_ERR_INVALID_ARG, "%s", _("Invalid argument")); - return -1; - } - if (maxnames == 0) { return 0; } -- 1.7.0.4

2 2

[libvirt] vcpupin: fix cpu affinity setting bug of qemu driver
by Taku Izumi 07 Jun '11

07 Jun '11

There is the case where cpu affinites for vcpu of qemu doesn't work correctly. For example, if only one vcpupin setting entry is provided and its setting is not for vcpu0, it doesn't work. # virsh dumpxml VM ... <vcpu>4</vcpu> <cputune> <vcpupin vcpu='3' cpuset='9-11'/> </cputune> ... # virsh start VM Domain VM started # virsh vcpuinfo VM VCPU: 0 CPU: 31 State: running CPU time: 2.5s CPU Affinity: yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy VCPU: 1 CPU: 12 State: running CPU time: 0.9s CPU Affinity: yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy VCPU: 2 CPU: 30 State: running CPU time: 1.5s CPU Affinity: yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy VCPU: 3 CPU: 13 State: running CPU time: 1.7s CPU Affinity: yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy This patch fixes this problem. Signed-off-by: Taku Izumi <izumi.taku(a)jp.fujitsu.com> --- src/qemu/qemu_process.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) Index: libvirt/src/qemu/qemu_process.c =================================================================== --- libvirt.orig/src/qemu/qemu_process.c +++ libvirt/src/qemu/qemu_process.c @@ -1194,7 +1194,7 @@ qemuProcessSetVcpuAffinites(virConnectPt virNodeInfo nodeinfo; pid_t vcpupid; unsigned char *cpumask; - int vcpu, cpumaplen, hostcpus, maxcpu; + int vcpu, cpumaplen, hostcpus, maxcpu, n; unsigned char *cpumap = NULL; int ret = -1; @@ -1223,14 +1223,12 @@ qemuProcessSetVcpuAffinites(virConnectPt return -1; } - for (vcpu = 0; vcpu < def->cputune.nvcpupin; vcpu++) { - if (vcpu != def->cputune.vcpupin[vcpu]->vcpuid) - continue; - + for (n = 0; n < def->cputune.nvcpupin; n++) { int i; + vcpu = def->cputune.vcpupin[n]->vcpuid; memset(cpumap, 0, cpumaplen); - cpumask = (unsigned char *)def->cputune.vcpupin[vcpu]->cpumask; + cpumask = (unsigned char *)def->cputune.vcpupin[n]->cpumask; vcpupid = priv->vcpupids[vcpu]; /* Convert cpumask to bitmap here. */

2 1

[libvirt] USB 2.0 support in qemu
by Gerd Hoffmann 07 Jun '11

07 Jun '11

Hi, Recently USB 2.0 support was merged into upstream qemu, I guess libvirt and virt-manager needs some bits to support this ... The docs/usb2.txt file in the qemu source tree carries some information on how to use it, and also some hints for USB pass-through. The most recent version (not yet merged upstream) can be found here: http://www.kraxel.org/cgit/qemu/tree/docs/usb2.txt?h=usb.15 If there are any questions feel free to ask. cheers, Gerd

2 1

[libvirt] [PATCH] docs: Add doc for video element
by Osier Yang 07 Jun '11

07 Jun '11

For backwards compatability, if no <video> is set but there is a <graphics> tag, then we add a default <video> according to the guest type. Add docs to tell the user about this to not make them confused. Especially if they remove the video (such as via "virsh edit"), it will be surprised for them to see the video element is still in domain XML. --- docs/formatdomain.html.in | 7 ++++++- 1 files changed, 6 insertions(+), 1 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index f8baffd..455f4dd 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1868,7 +1868,12 @@ qemu-kvm -net nic,model=? /dev/null <dt><code>video</code></dt> <dd> The <code>video</code> element is the a container for describing - video devices. + video devices. NB, for backwards compatability, if no <code>video</code> + is set but there is a <code>graphics</code> in domain xml, then libvirt + will add a default <code>video</code> according to the guest type, e.g. + For a guest of type "kvm", the default <code>video</code> for it is: + <code>type</code> with value "cirrus", <code>vram</code> with value "9216", + and <code>heads</code> with value "1". </dd> <dt><code>model</code></dt> -- 1.7.4

2 5

[libvirt] [PATCH] Avoid virGetVersion failure on specific driver support configurations
by Matthias Bolte 07 Jun '11

07 Jun '11

virGetVersion itself doesn't take a virConnectPtr, but in order to obtain the hypervisor version against which libvirt was compiled it is used in combination with virConnectGetType like this: hvType = virConnectGetType(conn) virGetVersion(&libVer, hvType, &typeVer) When virConnectGetType is called on a remote connection then the remote driver returns the type of the underlying driver on the server side, for example QEMU. Then virGetVersion compares hvType to a set of strings that depend on configure options and returns LIBVIR_VERSION_NUMBER in most cases. Now this fails in case libvirt on the client side is just compiled with the remote driver enabled only and the server side has the actual driver such as the QEMU driver. It just happens to work when the actual driver is enabled on client and server side. But that's not always true. I noticed this on FreeBSD: freebsd# virsh -c qemu+tcp://192.168.178.22/system version Compiled against library: libvir 0.9.2 error: failed to get the library version error: this function is not supported by the connection driver: virGetVersion This is not FreeBSD specific, happens on Windows as well due to the similar driver support configuration. The problem is that virConnectGetType returns QEMU, but virGetVersion on the client side only accepts Remote as hvType due to all other drivers being disabled on the client side. Daniel P. Berrange suggested to get rid of all the conditional code in virGetVersion, ignoring the hvType and always setting typeVer to LIBVIR_VERSION_NUMBER. virConnectGetVersion is supposed to be used to obtain the hypervisor version. --- src/libvirt.c | 86 ++++++-------------------------------------------------- 1 files changed, 10 insertions(+), 76 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index cbe1926..5a96625 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -894,21 +894,24 @@ int virStateActive(void) { /** * virGetVersion: * @libVer: return value for the library version (OUT) - * @type: the type of connection/driver looked at, if @typeVer is not NULL + * @type: the type of the connection driver, ignored, just pass NULL * @typeVer: optional return value for the version of the hypervisor (OUT) * * Provides information on up to two versions: @libVer is the version of the * library and will always be set unless an error occurs, in which case an - * error code will be returned. If @typeVer is not NULL it will be set to the - * version of the hypervisor @type against which the library was compiled. - * If @type is NULL, "Xen" is assumed, if @type is unknown or not - * available, an error code will be returned and @typeVer will be 0. + * error code will be returned. If @typeVer is not NULL it should have been be + * set to the version of the hypervisor @type against which the library was + * compiled. Due to a design problem this doesn't work as intented. Therefore, + * this function is only usefull to obtain the local library version number via + * @libVer. To get the version of the running hypervisor use the + * virConnectGetVersion function instead. To get the libvirt library version + * used by a connection use the virConnectGetLibVersion instead. * * Returns -1 in case of failure, 0 otherwise, and values for @libVer and * @typeVer have the format major * 1,000,000 + minor * 1,000 + release. */ int -virGetVersion(unsigned long *libVer, const char *type, +virGetVersion(unsigned long *libVer, const char *type ATTRIBUTE_UNUSED, unsigned long *typeVer) { VIR_DEBUG("libVir=%p, type=%s, typeVer=%p", libVer, type, typeVer); @@ -921,78 +924,9 @@ virGetVersion(unsigned long *libVer, const char *type, goto error; *libVer = LIBVIR_VERSION_NUMBER; - if (typeVer != NULL) { - if (type == NULL) - type = "Xen"; - -/* FIXME: Add _proper_ type version handling for loadable driver modules... */ -#ifdef WITH_DRIVER_MODULES + if (typeVer != NULL) *typeVer = LIBVIR_VERSION_NUMBER; -#else - *typeVer = 0; -# if WITH_XEN - if (STRCASEEQ(type, "Xen")) - *typeVer = xenUnifiedVersion(); -# endif -# if WITH_TEST - if (STRCASEEQ(type, "Test")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_QEMU - if (STRCASEEQ(type, "QEMU")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_LXC - if (STRCASEEQ(type, "LXC")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_LIBXL - if (STRCASEEQ(type, "xenlight")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_PHYP - if (STRCASEEQ(type, "phyp")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_OPENVZ - if (STRCASEEQ(type, "OpenVZ")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_VMWARE - if (STRCASEEQ(type, "VMware")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_VBOX - if (STRCASEEQ(type, "VBox")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_UML - if (STRCASEEQ(type, "UML")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_ONE - if (STRCASEEQ(type, "ONE")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_ESX - if (STRCASEEQ(type, "ESX")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_XENAPI - if (STRCASEEQ(type, "XenAPI")) - *typeVer = LIBVIR_VERSION_NUMBER; -# endif -# if WITH_REMOTE - if (STRCASEEQ(type, "Remote")) - *typeVer = remoteVersion(); -# endif - if (*typeVer == 0) { - virLibConnError(VIR_ERR_NO_SUPPORT, __FUNCTION__); - goto error; - } -#endif /* WITH_DRIVER_MODULES */ - } return 0; error: -- 1.7.0.4

2 2

[libvirt] [PATCH v2 0/3] add support for changing blkio parameters for inactive domains
by Hu Tao 06 Jun '11

06 Jun '11

This series enables user to change blkio parameters for inactive domains from virsh command line. CHANGES: v1-v2: - unify PARAM_CONFIG/PARAM_LIVE/PARAM_CURRENT Hu Tao (3): unify PARAM_CURRENT/PARAM_LIVE/PARAM_CONFIG flags update qemuDomainGetBlkioParameters to use flags Update qemuDomainSetBlkioParameters to use flags include/libvirt/libvirt.h.in | 58 ++---- src/esx/esx_driver.c | 10 +- src/libvirt.c | 4 +- src/openvz/openvz_driver.c | 10 +- src/qemu/qemu_driver.c | 442 ++++++++++++++++++++++++++---------------- src/test/test_driver.c | 42 ++-- src/uml/uml_driver.c | 4 +- src/vbox/vbox_tmpl.c | 22 +- tools/virsh.c | 74 ++++---- 9 files changed, 371 insertions(+), 295 deletions(-) -- 1.7.3.1

3 9