- Devel - libvirt List Archives

[libvirt] [PATCH 0/6] libxl bug and race fixes
by Jim Fehlig 25 Jan '13

25 Jan '13

This series fixes several bugs and races in the libxl driver, as described in the commit messages of the individual patches. One of the races was discussed on the xen-devel ml [1] and resulted in two fixes in libxl [2,3]. I've tested this series with and without the libxl patches. All of the previously discovered races are fixed with this series and the libxl patches. It is possible to fire asserts in libxl with this series and without the libxl patches, but that is no worse than the current behavior. The libxl patches have not yet been committed upstream, hence the 'XXXX' for commit id in patch 1. I will respond to the xen-devel thread now with an ACK on the libxl patches (and a reference to this series) so those can be committed. I'll fixup the commit message for patch 1 later, but wanted to get this series on the list for review before the 1.0.2 freeze. [1] http://lists.xen.org/archives/html/xen-devel/2012-11/msg01016.html [2] http://lists.xen.org/archives/html/xen-devel/2012-12/msg00684.html [3] http://lists.xen.org/archives/html/xen-devel/2012-12/msg00685.html Jim Fehlig (6): libxl: Fix handling of timeouts libxl: Fix races in libxl event code libxl: Fix race between destruction of objects libxl: Explicitly remove timeouts libxl: Fix removing non-persistent domain after save libxl: Find domain object in event handler src/libxl/libxl_conf.h | 9 ++ src/libxl/libxl_driver.c | 298 ++++++++++++++++++++++++++++++++++------------- 2 files changed, 226 insertions(+), 81 deletions(-) -- 1.8.0.1

2 18

[libvirt] [PATCH v2] python: Fix bindings for virDomainSnapshotGet{Domain, Connect}
by Jiri Denemark 25 Jan '13

25 Jan '13

https://bugzilla.redhat.com/show_bug.cgi?id=895882 virDomainSnapshot.getDomain() and virDomainSnapshot.getConnect() wrappers around virDomainSnapshotGet{Domain,Connect} were not supposed to be ever implemented. The class should contain proper domain() and connect() accessors that fetch python objects stored internally within the class. While domain() was already provided, connect() was missing. This patch adds connect() method to virDomainSnapshot class and reimplements getDomain() and getConnect() methods as aliases to domain() and connect() for backward compatibility. --- python/generator.py | 4 +++- python/libvirt-override-virDomainSnapshot.py | 8 ++++++++ src/libvirt.c | 10 ++++++++-- 3 files changed, 19 insertions(+), 3 deletions(-) diff --git a/python/generator.py b/python/generator.py index f853d77..5894616 100755 --- a/python/generator.py +++ b/python/generator.py @@ -527,6 +527,8 @@ skip_function = ( "virNWFilterGetConnect", "virStoragePoolGetConnect", "virStorageVolGetConnect", + "virDomainSnapshotGetConnect", + "virDomainSnapshotGetDomain", # only useful in C code, python code uses dict for typed parameters "virTypedParamsAddBoolean", @@ -953,7 +955,6 @@ classes_destructors = { class_skip_connect_impl = { "virConnect" : True, - "virDomainSnapshot": True, } class_domain_impl = { @@ -1436,6 +1437,7 @@ def buildWrappers(module): " self._conn = conn._conn\n") elif classname in [ "virDomainSnapshot" ]: classes.write(" self._dom = dom\n") + classes.write(" self._conn = dom.connect()\n") classes.write(" if _obj != None:self._o = _obj;return\n") classes.write(" self._o = None\n\n"); destruct=None diff --git a/python/libvirt-override-virDomainSnapshot.py b/python/libvirt-override-virDomainSnapshot.py index 3da7bfd..bf708a5 100644 --- a/python/libvirt-override-virDomainSnapshot.py +++ b/python/libvirt-override-virDomainSnapshot.py @@ -1,3 +1,11 @@ + def getConnect(self): + """Get the connection that owns the domain that a snapshot was created for""" + return self.connect() + + def getDomain(self): + """Get the domain that a snapshot was created for""" + return self.domain() + def listAllChildren(self, flags): """List all child snapshots and returns a list of snapshot objects""" ret = libvirtmod.virDomainSnapshotListAllChildren(self._o, flags) diff --git a/src/libvirt.c b/src/libvirt.c index d5d561c..fa39881 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -17850,7 +17850,10 @@ virDomainSnapshotGetName(virDomainSnapshotPtr snapshot) * virDomainSnapshotGetDomain: * @snapshot: a snapshot object * - * Get the domain that a snapshot was created for + * Get the domain that a snapshot was created for. + * + * WARNING: When writing libvirt bindings in other languages, do not use this + * function. Instead, store the domain and the snapshot object together. * * Returns the domain or NULL. */ @@ -17874,7 +17877,10 @@ virDomainSnapshotGetDomain(virDomainSnapshotPtr snapshot) * virDomainSnapshotGetConnect: * @snapshot: a snapshot object * - * Get the connection that owns the domain that a snapshot was created for + * Get the connection that owns the domain that a snapshot was created for. + * + * WARNING: When writing libvirt bindings in other languages, do not use this + * function. Instead, store the connection and the snapshot object together. * * Returns the connection or NULL. */ -- 1.8.1.1

3 5

[libvirt] Entering freeze for libvirt-1.0.2
by Daniel Veillard 25 Jan '13

25 Jan '13

As scheduled and to gte a release by the end of the month next week, I have just tagged a release candidate 1 version in git and pushed associated tarballs and rpms to the usual place: ftp://libvirt.org/libvirt/ please give it a try and let's ocuse on fixing bugs untl the freeze end. I will probably make a release candidate 2 on Monday, and if everything goes well rol out the 1.0.2 release on Wednesday or Thursday. thanks ! Daniel -- Daniel Veillard | Open Source and Standards, Red Hat veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

3 3

[libvirt] [PATCH] conf: avoid NULL deref for pmsuspended domain state
by Eric Blake 25 Jan '13

25 Jan '13

While working with a pmsuspend vs. snapshot issue, I noticed that the state file in /var/run/libvirt/qemu/dom.xml contained a rather suspicious "(null)" string, which does not round-trip well through a libvirtd restart. Had I been on a platform other than glibc where printf("%s",NULL) crashes instead of printing (null), we might have noticed the problem much sooner. And in fixing that problem, I also noticed that we had several missing states, because we were #defining several *_LAST names to a value _different_ than what they were already given as enums in libvirt.h. Yuck. I got rid of default: labels in the case statements, because they get in the way of gcc's -Wswitch helping us ensure we cover all enum values. * src/conf/domain_conf.c (virDomainStateReasonToString) (virDomainStateReasonFromString): Fill in missing domain states; rewrite case statement to let compiler enforce checking. (VIR_DOMAIN_NOSTATE_LAST, VIR_DOMAIN_RUNNING_LAST) (VIR_DOMAIN_BLOCKED_LAST, VIR_DOMAIN_PAUSED_LAST) (VIR_DOMAIN_SHUTDOWN_LAST, VIR_DOMAIN_SHUTOFF_LAST) (VIR_DOMAIN_CRASHED_LAST): Drop dead defines. (VIR_DOMAIN_PMSUSPENDED_LAST): Drop dead define. (virDomainPMSuspendedReason): Add missing enum function. (virDomainRunningReason, virDomainPausedReason): Add missing enum value. * src/conf/domain_conf.h (virDomainPMSuspendedReason): Declare missing functions. * src/libvirt_private.syms (domain_conf.h): Export them. --- src/conf/domain_conf.c | 32 +++++++++++++++++++------------- src/conf/domain_conf.h | 1 + src/libvirt_private.syms | 6 ++++-- 3 files changed, 24 insertions(+), 15 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 7273790..5782abb 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -574,11 +574,9 @@ VIR_ENUM_IMPL(virDomainState, VIR_DOMAIN_LAST, "crashed", "pmsuspended") -#define VIR_DOMAIN_NOSTATE_LAST (VIR_DOMAIN_NOSTATE_UNKNOWN + 1) VIR_ENUM_IMPL(virDomainNostateReason, VIR_DOMAIN_NOSTATE_LAST, "unknown") -#define VIR_DOMAIN_RUNNING_LAST (VIR_DOMAIN_RUNNING_SAVE_CANCELED + 1) VIR_ENUM_IMPL(virDomainRunningReason, VIR_DOMAIN_RUNNING_LAST, "unknown", "booted", @@ -587,13 +585,12 @@ VIR_ENUM_IMPL(virDomainRunningReason, VIR_DOMAIN_RUNNING_LAST, "from snapshot", "unpaused", "migration canceled", - "save canceled") + "save canceled", + "wakeup") -#define VIR_DOMAIN_BLOCKED_LAST (VIR_DOMAIN_BLOCKED_UNKNOWN + 1) VIR_ENUM_IMPL(virDomainBlockedReason, VIR_DOMAIN_BLOCKED_LAST, "unknown") -#define VIR_DOMAIN_PAUSED_LAST (VIR_DOMAIN_PAUSED_SHUTTING_DOWN + 1) VIR_ENUM_IMPL(virDomainPausedReason, VIR_DOMAIN_PAUSED_LAST, "unknown", "user", @@ -603,14 +600,13 @@ VIR_ENUM_IMPL(virDomainPausedReason, VIR_DOMAIN_PAUSED_LAST, "ioerror", "watchdog", "from snapshot", - "shutdown") + "shutdown", + "snapshot") -#define VIR_DOMAIN_SHUTDOWN_LAST (VIR_DOMAIN_SHUTDOWN_USER + 1) VIR_ENUM_IMPL(virDomainShutdownReason, VIR_DOMAIN_SHUTDOWN_LAST, "unknown", "user") -#define VIR_DOMAIN_SHUTOFF_LAST (VIR_DOMAIN_SHUTOFF_FROM_SNAPSHOT + 1) VIR_ENUM_IMPL(virDomainShutoffReason, VIR_DOMAIN_SHUTOFF_LAST, "unknown", "shutdown", @@ -621,10 +617,12 @@ VIR_ENUM_IMPL(virDomainShutoffReason, VIR_DOMAIN_SHUTOFF_LAST, "failed", "from snapshot") -#define VIR_DOMAIN_CRASHED_LAST (VIR_DOMAIN_CRASHED_UNKNOWN + 1) VIR_ENUM_IMPL(virDomainCrashedReason, VIR_DOMAIN_CRASHED_LAST, "unknown") +VIR_ENUM_IMPL(virDomainPMSuspendedReason, VIR_DOMAIN_PMSUSPENDED_LAST, + "unknown") + VIR_ENUM_IMPL(virDomainSeclabel, VIR_DOMAIN_SECLABEL_LAST, "default", "none", @@ -15438,9 +15436,13 @@ virDomainStateReasonToString(virDomainState state, int reason) return virDomainShutoffReasonTypeToString(reason); case VIR_DOMAIN_CRASHED: return virDomainCrashedReasonTypeToString(reason); - default: - return NULL; + case VIR_DOMAIN_PMSUSPENDED: + return virDomainPMSuspendedReasonTypeToString(reason); + case VIR_DOMAIN_LAST: + break; } + VIR_WARN("Unexpected domain state: %d", state); + return NULL; } @@ -15462,9 +15464,13 @@ virDomainStateReasonFromString(virDomainState state, const char *reason) return virDomainShutoffReasonTypeFromString(reason); case VIR_DOMAIN_CRASHED: return virDomainCrashedReasonTypeFromString(reason); - default: - return -1; + case VIR_DOMAIN_PMSUSPENDED: + return virDomainPMSuspendedReasonTypeFromString(reason); + case VIR_DOMAIN_LAST: + break; } + VIR_WARN("Unexpected domain state: %d", state); + return -1; } diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index abbfe86..9a9e0b1 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -2315,6 +2315,7 @@ VIR_ENUM_DECL(virDomainPausedReason) VIR_ENUM_DECL(virDomainShutdownReason) VIR_ENUM_DECL(virDomainShutoffReason) VIR_ENUM_DECL(virDomainCrashedReason) +VIR_ENUM_DECL(virDomainPMSuspendedReason) const char *virDomainStateReasonToString(virDomainState state, int reason); int virDomainStateReasonFromString(virDomainState state, const char *reason); diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index fc23adc..57ecc36 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -487,10 +487,12 @@ virDomainObjSetState; virDomainObjTaint; virDomainPausedReasonTypeFromString; virDomainPausedReasonTypeToString; -virDomainPciRombarModeTypeFromString; -virDomainPciRombarModeTypeToString; virDomainPMStateTypeFromString; virDomainPMStateTypeToString; +virDomainPMSuspendedReasonTypeFromString; +virDomainPMSuspendedReasonTypeToString; +virDomainPciRombarModeTypeFromString; +virDomainPciRombarModeTypeToString; virDomainRedirdevBusTypeFromString; virDomainRedirdevBusTypeToString; virDomainRemoveInactive; -- 1.8.1

3 4

[libvirt] [PATCH] Fix performance & reliabilty of QMP probing
by Daniel P. Berrange 25 Jan '13

25 Jan '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> This previous commit commit 1a50ba2cb07d8bb2aa724062889deb9efd7ad9e9 Author: Viktor Mihajlovski <mihajlov(a)linux.vnet.ibm.com> Date: Mon Nov 26 15:17:13 2012 +0100 qemu: Fix QMP Capabability Probing Failure which attempted to make sure the QEMU process used for probing ran as the right user id, caused serious performance regression and unreliability in probing. The -daemonize switch in QEMU guarantees that the monitor socket is present before the parent process exits. This means libvirtd is guaranteed to be able to connect immediately. By switching from -daemonize to the virCommandDaemonize API libvirtd was no longer synchronized with QEMU's startup process. The result was that the QEMU monitor failed to open and went into its 200ms sleep loop. This happened for all 25 binaries resulting in 5 seconds worth of sleeping at libvirtd startup. In addition sometimes when libvirt connected, QEMU would be partially initialized and crash causing total failure to probe that binary. This commit reverts the previous change, ensuring we do use the -daemonize flag to QEMU. Startup delay is cut from 7 seconds to 2 seconds on my machine, which is on a par with what it was prior to the capabilities rewrite. To deal with the fact that QEMU needs to be able to create the pidfile, we switch pidfile location fron runDir to libDir, which QEMU is guaranteed to be able to write to. Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/qemu/qemu_capabilities.c | 50 ++++++++++++++++++++++++++++++-------------- src/qemu/qemu_capabilities.h | 3 +-- 2 files changed, 35 insertions(+), 18 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 95fa3da..703179d 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -38,6 +38,7 @@ #include "virbitmap.h" #include "virnodesuspend.h" #include "qemu_monitor.h" +#include "virtime.h" #include <sys/stat.h> #include <unistd.h> @@ -916,11 +917,19 @@ virCapsPtr qemuCapsInit(qemuCapsCachePtr cache) * so just probe for them all - we gracefully fail * if a qemu-system-$ARCH binary can't be found */ - for (i = 0 ; i < VIR_ARCH_LAST ; i++) + unsigned long long a, b; + for (i = 0 ; i < VIR_ARCH_LAST ; i++) { + unsigned long long start, end; + if (virTimeMillisNow(&start) < 0) + goto error; if (qemuCapsInitGuest(caps, cache, virArchFromHost(), i) < 0) goto error; + if (virTimeMillisNow(&end) < 0) + goto error; + VIR_DEBUG("Probed %s in %llums", virArchToString(i), end-start); + } /* QEMU Requires an emulator in the XML */ virCapabilitiesSetEmulatorRequired(caps); @@ -2291,7 +2300,6 @@ qemuCapsInitQMPBasic(qemuCapsPtr caps) static int qemuCapsInitQMP(qemuCapsPtr caps, const char *libDir, - const char *runDir, uid_t runUid, gid_t runGid) { @@ -2324,8 +2332,11 @@ qemuCapsInitQMP(qemuCapsPtr caps, /* ".pidfile" suffix is used rather than ".pid" to avoid a possible clash * with a qemu domain called "capabilities" + * Normally we'd use runDir for pid files, but because we're using + * -daemonize we need QEMU to be allowed to create them, rather + * than libvirtd. So we're using libDir which QEMU can write to */ - if (virAsprintf(&pidfile, "%s/%s", runDir, "capabilities.pidfile") < 0) { + if (virAsprintf(&pidfile, "%s/%s", libDir, "capabilities.pidfile") < 0) { virReportOOMError(); goto cleanup; } @@ -2337,6 +2348,13 @@ qemuCapsInitQMP(qemuCapsPtr caps, VIR_DEBUG("Try to get caps via QMP caps=%p", caps); + /* + * We explicitly need to use -daemonize here, rather than + * virCommandDaemonize, because we need to synchronize + * with QEMU creating its monitor socket API. Using + * daemonize guarnatees control won't return to libvirt + * until the socket is present. + */ cmd = virCommandNewArgList(caps->binary, "-S", "-no-user-config", @@ -2344,14 +2362,14 @@ qemuCapsInitQMP(qemuCapsPtr caps, "-nographic", "-M", "none", "-qmp", monarg, + "-pidfile", pidfile, + "-daemonize", NULL); virCommandAddEnvPassCommon(cmd); virCommandClearCaps(cmd); hookData.runUid = runUid; hookData.runGid = runGid; virCommandSetPreExecHook(cmd, qemuCapsHook, &hookData); - virCommandSetPidFile(cmd, pidfile); - virCommandDaemonize(cmd); if (virCommandRun(cmd, &status) < 0) goto cleanup; @@ -2472,7 +2490,6 @@ cleanup: qemuCapsPtr qemuCapsNewForBinary(const char *binary, const char *libDir, - const char *runDir, uid_t runUid, gid_t runGid) { @@ -2502,12 +2519,14 @@ qemuCapsPtr qemuCapsNewForBinary(const char *binary, goto error; } - if ((rv = qemuCapsInitQMP(caps, libDir, runDir, runUid, runGid)) < 0) + if ((rv = qemuCapsInitQMP(caps, libDir, runUid, runGid)) < 0) goto error; - if (!caps->usedQMP && - qemuCapsInitHelp(caps, runUid, runGid) < 0) - goto error; + if (!caps->usedQMP) { + VIR_ERROR("Falling back to help"); + if (qemuCapsInitHelp(caps, runUid, runGid) < 0) + goto error; + } return caps; @@ -2542,8 +2561,9 @@ qemuCapsHashDataFree(void *payload, const void *key ATTRIBUTE_UNUSED) qemuCapsCachePtr -qemuCapsCacheNew(const char *libDir, const char *runDir, - uid_t runUid, gid_t runGid) +qemuCapsCacheNew(const char *libDir, + uid_t runUid, + gid_t runGid) { qemuCapsCachePtr cache; @@ -2561,8 +2581,7 @@ qemuCapsCacheNew(const char *libDir, const char *runDir, if (!(cache->binaries = virHashCreate(10, qemuCapsHashDataFree))) goto error; - if (!(cache->libDir = strdup(libDir)) || - !(cache->runDir = strdup(runDir))) { + if (!(cache->libDir = strdup(libDir))) { virReportOOMError(); goto error; } @@ -2594,7 +2613,7 @@ qemuCapsCacheLookup(qemuCapsCachePtr cache, const char *binary) if (!ret) { VIR_DEBUG("Creating capabilities for %s", binary); - ret = qemuCapsNewForBinary(binary, cache->libDir, cache->runDir, + ret = qemuCapsNewForBinary(binary, cache->libDir, cache->runUid, cache->runGid); if (ret) { VIR_DEBUG("Caching capabilities %p for %s", @@ -2634,7 +2653,6 @@ qemuCapsCacheFree(qemuCapsCachePtr cache) return; VIR_FREE(cache->libDir); - VIR_FREE(cache->runDir); virHashFree(cache->binaries); virMutexDestroy(&cache->lock); VIR_FREE(cache); diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index 089fa30..5279d07 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -179,7 +179,6 @@ qemuCapsPtr qemuCapsNew(void); qemuCapsPtr qemuCapsNewCopy(qemuCapsPtr caps); qemuCapsPtr qemuCapsNewForBinary(const char *binary, const char *libDir, - const char *runDir, uid_t runUid, gid_t runGid); @@ -219,7 +218,7 @@ int qemuCapsGetMachineTypesCaps(qemuCapsPtr caps, bool qemuCapsIsValid(qemuCapsPtr caps); -qemuCapsCachePtr qemuCapsCacheNew(const char *libDir, const char *runDir, +qemuCapsCachePtr qemuCapsCacheNew(const char *libDir, uid_t uid, gid_t gid); qemuCapsPtr qemuCapsCacheLookup(qemuCapsCachePtr cache, const char *binary); qemuCapsPtr qemuCapsCacheLookupCopy(qemuCapsCachePtr cache, const char *binary); -- 1.8.0.2

4 6

[libvirt] [Openstack][Sheepdog][Libvirt][Qemu]Add a new block storage driver by Libvirt/Qemu way for Openstack
by harryxiyou 25 Jan '13

25 Jan '13

Hi all, I wanna add a new block storage driver by Libvirt/Qemu way for Openstack, which is as same as Sheepdog driver for Openstack. So i think the theories are like this. 1, In the Openstack Nova branch, Openstck driver call libvirt client and send parameters to libvirt client.(From this point, i should modify Openstack Nova source codes. They are a, nova/nova/virt/libvirt/driver.py add new driver way b, /OpenStack/nova/nova/tests/test_libvirt.py add new driver test) 2, According to own protocol, libvirt client in Openstack Nova branch send parameters to Libvirt server.(From this point, i should modify libvirt library to let libvirt library support this new driver like Sheepdog). 3, Libvirt server call Qemu interfaces to send parameters to Qemu.(From this point, i should modify Qemu source codes to let Qemu support this new driver like Sheepdog). 4, In Openstack Cinder branch, Openstack driver use Qemu commands to create this new volumes to Qemu.(From this point, i should modify Openstack Cinder branch source codes like this. a, Add new driver file /OpenStack/cinder/cinder/volume/drivers/new_driver.py like Sheepdog.py b, Change file /OpenStack/cinder/cinder/tests/test_drivers_compatibility.py to test new driver). 5, At last, i should also modify /OpenStack/manuals/doc/src/docbkx/openstack-compute-admin/tables/hypervisors-nova-conf.xml to configure this new driver. Are my theories right? Should i do any other stuffs? Could anyone give me any other suggestions? Thanks in advance ;-) -- Thanks Harry Wei

2 5

[libvirt] [PATCH] virsh: Forbid usage of --memspec and --disk-only together
by Peter Krempa 25 Jan '13

25 Jan '13

virsh snapshot-create-as with --memspec and --disk only fails at the qemu driver. Forbid the creation at virsh too and document it. --- tools/virsh-snapshot.c | 6 ++++++ tools/virsh.pod | 3 ++- 2 files changed, 8 insertions(+), 1 deletion(-) diff --git a/tools/virsh-snapshot.c b/tools/virsh-snapshot.c index e4745d6..86fd1fd 100644 --- a/tools/virsh-snapshot.c +++ b/tools/virsh-snapshot.c @@ -457,6 +457,12 @@ cmdSnapshotCreateAs(vshControl *ctl, const vshCmd *cmd) goto cleanup; } + if (memspec && (flags & VIR_DOMAIN_SNAPSHOT_CREATE_DISK_ONLY)) { + vshError(ctl, "%s", + _("Can't combine disk only snapshots with --memspec")); + goto cleanup; + } + if (memspec && vshParseSnapshotMemspec(ctl, &buf, memspec) < 0) goto cleanup; diff --git a/tools/virsh.pod b/tools/virsh.pod index e2a2aec..c5cf16c 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -2746,7 +2746,8 @@ The I<--memspec> option can be used to control whether a checkpoint is internal or external. The I<--memspec> flag is mandatory, followed by a B<memspec> of the form B<[file=]name[,snapshot=type]>, where type can be B<none>, B<internal>, or B<external>. To include a literal -comma in B<file=name>, escape it with a second comma. +comma in B<file=name>, escape it with a second comma. I<--memspec> cannot +be used together with I<--disk-only>. The I<--diskspec> option can be used to control how I<--disk-only> and external checkpoints create external files. This option can occur -- 1.8.1.1

2 1

[libvirt] [PATCH] Fix bogus reporting of KVM support for non-native emulators
by Daniel P. Berrange 25 Jan '13

25 Jan '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> A logic bug meant we reported KVM was possible for every architecture, merely based on whether the query-kvm command exists. We should instead have been doing it based on whether the query-kvm command returns 'present: 1' Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/qemu/qemu_capabilities.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index f86c28f..29693c3 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -2117,11 +2117,16 @@ qemuCapsProbeQMPKVMState(qemuCapsPtr caps, return -1; /* The QEMU_CAPS_KVM flag was initially set according to the QEMU - * reporting the recognition of 'query-kvm' QMP command, but the - * flag means whether the KVM is enabled by default and should be - * disabled in case we want SW emulated machine, so let's fix that - * if it's true. */ - if (!enabled) { + * reporting the recognition of 'query-kvm' QMP command. That merely + * indicates existance of the command though, not whether KVM support + * is actually available, nor whether it is enabled by default. + * + * If it is not present we need to clear the flag, and if it is + * not enabled by default we need to change the flag. + */ + if (!present) { + qemuCapsClear(caps, QEMU_CAPS_KVM); + } else if (!enabled) { qemuCapsClear(caps, QEMU_CAPS_KVM); qemuCapsSet(caps, QEMU_CAPS_ENABLE_KVM); } -- 1.8.0.2

2 1

[libvirt] [PATCH] qemu: don't share kerberos caches between domains
by Eric Blake 25 Jan '13

25 Jan '13

https://bugzilla.redhat.com/show_bug.cgi?id=718377 complains that there were some SELinux AVCs when using vnc console over Kerberos. The root problem was that Kerberos tries to set up a cache file, and if we don't tell it where, then all domains use the same cache file, which violates sVirt protections. Setting the environment variable unconditionally should be safe, even for setups where Kerboros won't actually create a cache file. * src/qemu/qemu_process.c (qemuProcessStart): Set KRB5CACHEDIR for each domain. --- src/qemu/qemu_process.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 55d00e3..986e8ee 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -1,7 +1,7 @@ /* * qemu_process.h: QEMU process management * - * Copyright (C) 2006-2012 Red Hat, Inc. + * Copyright (C) 2006-2013 Red Hat, Inc. * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -3857,6 +3857,8 @@ int qemuProcessStart(virConnectPtr conn, virCommandSetPidFile(cmd, priv->pidfile); virCommandDaemonize(cmd); virCommandRequireHandshake(cmd); + virCommandAddEnvFormat(cmd, "KRB5CACHEDIR=%s/%s.krb", + driver->cacheDir, vm->def->name); ret = virCommandRun(cmd, NULL); -- 1.8.1

4 6

[libvirt] [PATCH] qemuDomainSendKey: Relax the qemu driver locking
by Michal Privoznik 25 Jan '13

25 Jan '13

Currently, there is no reason to hold qemu driver locked throughout whole API execution. Moreover, we can use the new qemuDomObjFromDomain() internal API to lookup domain then. --- src/qemu/qemu_driver.c | 16 ++++------------ 1 file changed, 4 insertions(+), 12 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 6d4c1e9..100f10b 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2407,19 +2407,12 @@ static int qemuDomainSendKey(virDomainPtr domain, } } - qemuDriverLock(driver); - vm = virDomainFindByUUID(&driver->domains, domain->uuid); - if (!vm) { - char uuidstr[VIR_UUID_STRING_BUFLEN]; - virUUIDFormat(domain->uuid, uuidstr); - virReportError(VIR_ERR_NO_DOMAIN, - _("no domain with matching uuid '%s'"), uuidstr); + if (!(vm = qemuDomObjFromDomain(domain))) goto cleanup; - } priv = vm->privateData; - if (qemuDomainObjBeginJobWithDriver(driver, vm, QEMU_JOB_MODIFY) < 0) + if (qemuDomainObjBeginJob(driver, vm, QEMU_JOB_MODIFY) < 0) goto cleanup; if (!virDomainObjIsActive(vm)) { @@ -2428,9 +2421,9 @@ static int qemuDomainSendKey(virDomainPtr domain, goto endjob; } - qemuDomainObjEnterMonitorWithDriver(driver, vm); + qemuDomainObjEnterMonitor(driver, vm); ret = qemuMonitorSendKey(priv->mon, holdtime, keycodes, nkeycodes); - qemuDomainObjExitMonitorWithDriver(driver, vm); + qemuDomainObjExitMonitor(driver, vm); endjob: if (qemuDomainObjEndJob(driver, vm) == 0) @@ -2439,7 +2432,6 @@ endjob: cleanup: if (vm) virObjectUnlock(vm); - qemuDriverUnlock(driver); return ret; } -- 1.8.0.2

5 7