- Devel - libvirt List Archives

[libvirt] [PATCH] docs: network: fix some trivial typos in docs/formatnetwork.html
by Chen Fan 06 Dec '14

06 Dec '14

this patch fix some weird typos: 1. < hostdev> => <hostdev> 2. < type> => <type> 3. <virtualport > => <virtualport> 4. redundant comma 5. missing right-half bracket Signed-off-by: Chen Fan <chen.fan.fnst(a)cn.fujitsu.com> --- docs/formatnetwork.html.in | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/docs/formatnetwork.html.in b/docs/formatnetwork.html.in index dc438ae..7bcf316 100644 --- a/docs/formatnetwork.html.in +++ b/docs/formatnetwork.html.in @@ -288,7 +288,7 @@ (Single Root I/O Virtualization) virtual function (VF) devices can be assigned in this manner; to assign a standard single-port PCI or PCIe ethernet card to a guest, - use the traditional <code>< hostdev></code> device + use the traditional <code><hostdev></code> device definition. Since 0.10.0 @@ -312,9 +312,9 @@ Note that this "intelligent passthrough" of network devices is very similar to the functionality of a - standard <code>< hostdev></code> device, the + standard <code><hostdev></code> device, the difference being that this method allows specifying a MAC - address, vlan tag, and <code><virtualport ></code> + address, vlan tag, and <code><virtualport></code> for the passed-through device. If these capabilities are not required, if you have a standard single-port PCI, PCIe, or USB network card that doesn't support SR-IOV (and @@ -383,9 +383,9 @@ since 0.10.0 When using forward mode 'hostdev', the interface pool is specified with a list of <code><address></code> elements, each of which has - <code>< type></code> (must always be <code>'pci'</code>, + <code><type></code> (must always be <code>'pci'</code>), <code><domain></code>, <code><bus></code>, - <code><slot></code>, and <code><function></code> + <code><slot></code>and <code><function></code> attributes. <pre> -- 1.9.3

2 1

[libvirt] [PATCH 0/4] Add auth and adapter to virsh pool-{create|define}-as
by John Ferlan 06 Dec '14

06 Dec '14

Recently while reviewing some documentation for iSCSI authentication description I found that virsh pool-{create|define}-as did not provide a mechanism to add the <auth> and/or <adapter> options on the command line - so I figured adding them would be simple. In the course of doing this I also noted a few typos on the formatstorage html page and (to say the last) a fairly incomplete or lacking man page description for pool-{create|define}-as. John Ferlan (4): docs: Fix a couple of typos on the storage pool html virsh.pod: Fix the pool-define-as and pool-create-as description virsh: Add auth options for pool-{create|define}-as virsh: Add adapter options for pool-{create|define}-as docs/formatstorage.html.in | 4 ++-- tools/virsh-pool.c | 59 ++++++++++++++++++++++++++++++++++++++++++++-- tools/virsh.pod | 53 ++++++++++++++++++++++++++++++++++++----- 3 files changed, 106 insertions(+), 10 deletions(-) -- 1.9.3

3 7

[libvirt] [PATCH v2] automatic create tap device with network type ethernet
by Vasiliy Tolstov 05 Dec '14

05 Dec '14

If user not specify script in network type ethernet, assume that user needs simple tap device created with libvirt. This patch does not need to run external script to create tap device or add root to qemu process. Signed-off-by: Vasiliy Tolstov <v.tolstov(a)selfip.ru> --- src/qemu/qemu_command.c | 86 ++++++++++++++++++++++++++++++++----------------- src/qemu/qemu_hotplug.c | 10 ++---- src/qemu/qemu_process.c | 4 +++ 3 files changed, 62 insertions(+), 38 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 4ed6506..0911c32 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -319,7 +319,7 @@ qemuNetworkIfaceConnect(virDomainDefPtr def, } else if (actualType == VIR_DOMAIN_NET_TYPE_BRIDGE) { if (VIR_STRDUP(brname, virDomainNetGetActualBridgeName(net)) < 0) return ret; - } else { + } else if (actualType != VIR_DOMAIN_NET_TYPE_ETHERNET) { virReportError(VIR_ERR_INTERNAL_ERROR, _("Network type %d is not supported"), virDomainNetGetActualType(net)); @@ -341,30 +341,40 @@ qemuNetworkIfaceConnect(virDomainDefPtr def, tap_create_flags |= VIR_NETDEV_TAP_CREATE_VNET_HDR; } - if (cfg->privileged) { - if (virNetDevTapCreateInBridgePort(brname, &net->ifname, &net->mac, - def->uuid, tunpath, tapfd, *tapfdSize, - virDomainNetGetActualVirtPortProfile(net), - virDomainNetGetActualVlan(net), - tap_create_flags) < 0) { + if (actualType == VIR_DOMAIN_NET_TYPE_ETHERNET) { + if (virNetDevTapCreate(&net->ifname, tunpath, tapfd, *tapfdSize, + tap_create_flags) < 0) { virDomainAuditNetDevice(def, net, tunpath, false); goto cleanup; } - } else { - if (qemuCreateInBridgePortWithHelper(cfg, brname, - &net->ifname, - tapfd, tap_create_flags) < 0) { - virDomainAuditNetDevice(def, net, tunpath, false); + if (virNetDevSetOnline(net->ifname, !!(tap_create_flags & VIR_NETDEV_TAP_CREATE_IFUP)) < 0) goto cleanup; + } else { + if (cfg->privileged) { + if (virNetDevTapCreateInBridgePort(brname, &net->ifname, &net->mac, + def->uuid, tunpath, tapfd, *tapfdSize, + virDomainNetGetActualVirtPortProfile(net), + virDomainNetGetActualVlan(net), + tap_create_flags) < 0) { + virDomainAuditNetDevice(def, net, tunpath, false); + goto cleanup; + } + } else { + if (qemuCreateInBridgePortWithHelper(cfg, brname, + &net->ifname, + tapfd, tap_create_flags) < 0) { + virDomainAuditNetDevice(def, net, tunpath, false); + goto cleanup; + } + /* qemuCreateInBridgePortWithHelper can only create a single FD */ + if (*tapfdSize > 1) { + VIR_WARN("Ignoring multiqueue network request"); + *tapfdSize = 1; + } } - /* qemuCreateInBridgePortWithHelper can only create a single FD */ - if (*tapfdSize > 1) { - VIR_WARN("Ignoring multiqueue network request"); - *tapfdSize = 1; - } - } - virDomainAuditNetDevice(def, net, tunpath, true); + virDomainAuditNetDevice(def, net, tunpath, true); + } if (cfg->macFilter && ebtablesAddForwardAllowIn(driver->ebtables, @@ -4540,18 +4550,32 @@ qemuBuildHostNetStr(virDomainNetDefPtr net, break; case VIR_DOMAIN_NET_TYPE_ETHERNET: - virBufferAddLit(&buf, "tap"); + virBufferAddLit(&buf, "tap"); + type_sep = ','; + if (net->script) { if (net->ifname) { - virBufferAsprintf(&buf, "%cifname=%s", type_sep, net->ifname); - type_sep = ','; + virBufferAsprintf(&buf, "%cifname=%s", type_sep, net->ifname); + type_sep = ','; } - if (net->script) { - virBufferAsprintf(&buf, "%cscript=%s", type_sep, - net->script); - type_sep = ','; + virBufferAsprintf(&buf, "%cscript=%s", type_sep, net->script); + type_sep = ','; + } else { + /* for one tapfd 'fd=' shall be used, + * for more than one 'fds=' is the right choice */ + if (tapfdSize == 1) { + virBufferAsprintf(&buf, "%cfd=%s", type_sep, tapfd[0]); + } else { + virBufferAsprintf(&buf, "%cfds=", type_sep); + for (i = 0; i < tapfdSize; i++) { + if (i) + virBufferAddChar(&buf, ':'); + virBufferAdd(&buf, tapfd[i], -1); + } } - is_tap = true; - break; + type_sep = ','; + } + is_tap = true; + break; case VIR_DOMAIN_NET_TYPE_CLIENT: virBufferAsprintf(&buf, "socket%cconnect=%s:%d", @@ -7372,7 +7396,8 @@ qemuBuildInterfaceCommandLine(virCommandPtr cmd, /* Currently nothing besides TAP devices supports multiqueue. */ if (net->driver.virtio.queues > 0 && !(actualType == VIR_DOMAIN_NET_TYPE_NETWORK || - actualType == VIR_DOMAIN_NET_TYPE_BRIDGE)) { + actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET)) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, _("Multiqueue network is not supported for: %s"), virDomainNetTypeToString(actualType)); @@ -7380,7 +7405,8 @@ qemuBuildInterfaceCommandLine(virCommandPtr cmd, } if (actualType == VIR_DOMAIN_NET_TYPE_NETWORK || - actualType == VIR_DOMAIN_NET_TYPE_BRIDGE) { + actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET) { tapfdSize = net->driver.virtio.queues; if (!tapfdSize) tapfdSize = 1; diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index b00fd8f..bf92e4d 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -908,7 +908,8 @@ int qemuDomainAttachNetDevice(virConnectPtr conn, } if (actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || - actualType == VIR_DOMAIN_NET_TYPE_NETWORK) { + actualType == VIR_DOMAIN_NET_TYPE_NETWORK || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET) { tapfdSize = vhostfdSize = net->driver.virtio.queues; if (!tapfdSize) tapfdSize = vhostfdSize = 1; @@ -939,13 +940,6 @@ int qemuDomainAttachNetDevice(virConnectPtr conn, iface_connected = true; if (qemuOpenVhostNet(vm->def, net, priv->qemuCaps, vhostfd, &vhostfdSize) < 0) goto cleanup; - } else if (actualType == VIR_DOMAIN_NET_TYPE_ETHERNET) { - vhostfdSize = 1; - if (VIR_ALLOC(vhostfd) < 0) - goto cleanup; - *vhostfd = -1; - if (qemuOpenVhostNet(vm->def, net, priv->qemuCaps, vhostfd, &vhostfdSize) < 0) - goto cleanup; } /* Set Bandwidth */ diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 382d802..f161e3a 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -5080,6 +5080,10 @@ void qemuProcessStop(virQEMUDriverPtr driver, cfg->stateDir)); VIR_FREE(net->ifname); break; + case VIR_DOMAIN_NET_TYPE_ETHERNET: + ignore_value(virNetDevTapDelete(net->ifname, net->backend.tap)); + VIR_FREE(net->ifname); + break; case VIR_DOMAIN_NET_TYPE_BRIDGE: case VIR_DOMAIN_NET_TYPE_NETWORK: #ifdef VIR_NETDEV_TAP_REQUIRE_MANUAL_CLEANUP -- 2.1.3

2 5

[libvirt] [PATCH] qemu: snapshot: Report better error when create internal snapshot with passthrough devices
by Shanzhi Yu 05 Dec '14

05 Dec '14

When create internal system checkpoint snapshot, it will not works if guest has passthrough devices attached. It will report error: error: operation failed: Error -22 while writing VM With this patch, it will report erro: error: Requested operation is not valid: domain has assigned non-USB host devices Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=874418#c19 --- src/qemu/qemu_driver.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 5dc62b0..0fdee26 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -12902,6 +12902,9 @@ qemuDomainSnapshotCreateActiveInternal(virConnectPtr conn, } } + if (!qemuMigrationIsAllowed(driver, vm, vm->def, false, false)) + goto cleanup; + if (qemuDomainObjEnterMonitorAsync(driver, vm, QEMU_ASYNC_JOB_SNAPSHOT) < 0) { resume = false; -- 1.8.3.1

2 1

[libvirt] question about MigrateSetMaxDowntime
by Jim Fehlig 05 Dec '14

05 Dec '14

A user asked me about setting max downtime prior to migration, which I thought could be done similar to MigrateSetMaxSpeed. I was surprised to find that it can only be set when migration is in progress # virsh migrate-setmaxdowntime test 1000 error: Requested operation is not valid: domain is not being migrated Is there any reason this cannot be set prior to migration? Max speed can even be set on a persistent domain that is not running. I'd be happy to work on a patch if the only reason is no one has needed this itch scratched. Regards, Jim

2 2

[libvirt] [PATCH] storage: backend: Log uid/gid when initializing storage file backend
by Peter Krempa 05 Dec '14

05 Dec '14

To ease debugging permission problems add uid/gid values to the debug message when initializing a storage file backend. --- src/storage/storage_backend_fs.c | 4 ++-- src/storage/storage_backend_gluster.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/src/storage/storage_backend_fs.c b/src/storage/storage_backend_fs.c index 0ee1d09..b831268 100644 --- a/src/storage/storage_backend_fs.c +++ b/src/storage/storage_backend_fs.c @@ -1370,9 +1370,9 @@ virStorageFileBackendFileInit(virStorageSourcePtr src) { virStorageFileBackendFsPrivPtr priv = NULL; - VIR_DEBUG("initializing FS storage file %p (%s:%s)", src, + VIR_DEBUG("initializing FS storage file %p (%s:%s[%u:%u])", src, virStorageTypeToString(virStorageSourceGetActualType(src)), - src->path); + src->path, src->drv->uid, src->drv->gid); if (VIR_ALLOC(priv) < 0) return -1; diff --git a/src/storage/storage_backend_gluster.c b/src/storage/storage_backend_gluster.c index b79b634..9d7b0f5 100644 --- a/src/storage/storage_backend_gluster.c +++ b/src/storage/storage_backend_gluster.c @@ -582,9 +582,9 @@ virStorageFileBackendGlusterInit(virStorageSourcePtr src) hostname = host->name; - VIR_DEBUG("initializing gluster storage file %p (gluster://%s:%s/%s%s)", + VIR_DEBUG("initializing gluster storage file %p (gluster://%s:%s/%s%s)[%u:%u]", src, hostname, host->port ? host->port : "0", - NULLSTR(src->volume), src->path); + NULLSTR(src->volume), src->path, src->drv->uid, src->drv->gid); if (!src->volume) { virReportError(VIR_ERR_INTERNAL_ERROR, -- 2.1.0

2 2

[libvirt] [PATCHv2 2/2] security: Add a new func use stat to get process DAC label
by Luyao Huang 05 Dec '14

05 Dec '14

When use qemuProcessAttach to attach a qemu process, cannot get a right DAC label. Add a new func to get process label via stat func. Do not remove virDomainDefGetSecurityLabelDef before try to use stat to get process DAC label, because There are some other func call virSecurityDACGetProcessLabel. v2 add support freeBSD. Signed-off-by: Luyao Huang <lhuang(a)redhat.com> --- src/security/security_dac.c | 95 ++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 93 insertions(+), 2 deletions(-) diff --git a/src/security/security_dac.c b/src/security/security_dac.c index 85253af..89cafa3 100644 --- a/src/security/security_dac.c +++ b/src/security/security_dac.c @@ -23,6 +23,11 @@ #include <sys/stat.h> #include <fcntl.h> +#ifdef __FreeBSD__ +# include <sys/sysctl.h> +# include <sys/user.h> +#endif + #include "security_dac.h" #include "virerror.h" #include "virfile.h" @@ -1236,18 +1241,104 @@ virSecurityDACReserveLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED, return 0; } +#ifdef __linux__ +static int +virSecurityDACGetProcessLabelInternal(pid_t pid, + virSecurityLabelPtr seclabel) +{ + struct stat sb; + char *path = NULL; + char *label = NULL; + int ret = -1; + + VIR_INFO("Getting DAC user and group on process '%d'", pid); + + if (virAsprintf(&path, "/proc/%d", (int) pid) < 0) + goto cleanup; + + if (lstat(path, &sb) < 0) + goto cleanup; + + if (virAsprintf(&label, "+%u:+%u", + (unsigned int) sb.st_uid, + (unsigned int) sb.st_gid) < 0) + goto cleanup; + + if (virStrcpy(seclabel->label, label,VIR_SECURITY_LABEL_BUFLEN) == NULL) + goto cleanup; + ret = 0; + +cleanup: + VIR_FREE(path); + VIR_FREE(label); + return ret; +} +#elif defined(__FreeBSD__) +static int +virSecurityDACGetProcessLabelInternal(pid_t pid, + virSecurityLabelPtr seclabel) +{ + struct kinfo_proc p; + int mib[4]; + size_t len = 4; + char *label = NULL; + int ret = -1; + + sysctlnametomib("kern.proc.pid", mib, &len); + + len = sizeof(struct kinfo_proc); + mib[3] = pid; + + if (sysctl(mib, 4, &p, &len, NULL, 0) < 0) + goto cleanup; + + if (virAsprintf(&label, "+%u:+%u", + (unsigned int) p.ki_ruid, + (unsigned int) p.ki_rgid) < 0) + goto cleanup; + + if (virStrcpy(seclabel->label, label,VIR_SECURITY_LABEL_BUFLEN) == NULL) + goto cleanup; + ret = 0; + +cleanup: + VIR_FREE(label); + return ret; +} +#else +static int +virSecurityDACGetProcessLabelInternal(pid_t pid, + virSecurityLabelPtr seclabel) +{ + return -1; +} +#endif + static int virSecurityDACGetProcessLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED, virDomainDefPtr def, - pid_t pid ATTRIBUTE_UNUSED, + pid_t pid, virSecurityLabelPtr seclabel) { virSecurityLabelDefPtr secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_DAC_NAME); - if (!secdef || !seclabel) + if (!seclabel) return -1; + if (secdef == NULL) { + VIR_DEBUG("missing label for DAC security " + "driver in domain %s", def->name); + + if (virSecurityDACGetProcessLabelInternal(pid, seclabel) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Cannot get process %d DAC label"),pid); + return -1; + } + + return 0; + } + if (secdef->label) ignore_value(virStrcpy(seclabel->label, secdef->label, VIR_SECURITY_LABEL_BUFLEN)); -- 1.8.3.1

3 2

[libvirt] [PATCH] networkValidate: Disallow bandwidth in portgroups too
by Michal Privoznik 05 Dec '14

05 Dec '14

In one of the previous commits (eafb53fe) we disallowed network-wide bandwidth to some network types. However, we forgot about <portgroups/> which can have <bandwidth/> too. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/network/bridge_driver.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/src/network/bridge_driver.c b/src/network/bridge_driver.c index 9ccc9f8..1459f04 100644 --- a/src/network/bridge_driver.c +++ b/src/network/bridge_driver.c @@ -2711,6 +2711,7 @@ networkValidate(virNetworkDefPtr def, virPortGroupDefPtr defaultPortGroup = NULL; virNetworkIpDefPtr ipdef; bool ipv4def = false, ipv6def = false; + bool bandwidthAllowed = true; /* check for duplicate networks */ if (virNetworkObjIsDuplicate(&driver->networks, def, check_active) < 0) @@ -2772,6 +2773,7 @@ networkValidate(virNetworkDefPtr def, virNetworkForwardTypeToString(def->forward.type)); return -1; } + bandwidthAllowed = false; } /* We only support dhcp on one IPv4 address and @@ -2850,6 +2852,15 @@ networkValidate(virNetworkDefPtr def, } defaultPortGroup = &def->portGroups[i]; } + + if (def->portGroups[i].bandwidth && !bandwidthAllowed) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Unsupported <bandwidth> element in network '%s' " + "in portgroup '%s' with forward mode='%s'"), + def->name, def->portGroups[i].name, + virNetworkForwardTypeToString(def->forward.type)); + return -1; + } } if (badVlanUse || (vlanUsed && !vlanAllowed && !defaultPortGroup)) { -- 2.0.4

2 1

[libvirt] [PATCH 00/11] Post-Copy Live Migration Support
by Cristian Klein 05 Dec '14

05 Dec '14

Qemu currently implements pre-copy live migration. VM memory pages are first copied from the source hypervisor to the destination, potentially multiple times as pages get dirtied during transfer, then VCPU state is migrated. Unfortunately, if the VM dirties memory faster than the network bandwidth, then pre-copy cannot finish. `virsh` currently includes an option to suspend a VM after a timeout, so that migration may finish, but at the expense of downtime. A future version of qemu will implement post-copy live migration. The VCPU state is first migrated to the destination hypervisor, then memory pages are pulled from the source hypervisor. Post-copy has the potential to do migration with zero-downtime, despite the VM dirtying pages fast, with minimum performance impact. On the other hand, while post-copy is in progress, any network failure would render the VM unusable, as its memory is partitioned between the source and destination hypervisor. Therefore, post-copy should only be used when necessary. Post-copy migration in qemu will work as follows: (1) The `x-postcopy-ram` migration capability needs to be set. (2) Migration is started. (3) When the user decides so, post-copy migration is activated by sending the `migrate-start-postcopy` command. (4) Qemu acknowledges by setting migration status to `postcopy-active`. This patch series implements two ways to access post-copy functionality: low-level and high-level. The low-level API implements a mechanism that basically requires the libvirt user to manually go through the above steps, by calling migration with `VIR_MIGRATE_ENABLE_POSTCOPY`, then during migration from a separate thread, call `virDomainMigrateStartPostCopy`. The choice of when migration should switch from pre-copy to post-copy is left entirely to the user. The high-level API implements a policy that automatically triggers post-copy after one pass of pre-copy, which experiments have shown to minimize downtime. Using it is also simpler: the user only has to call migration with `VIR_MIGRATE_POSTCOPY_AFTER_PRECOPY`. TODO: - Wait for qemu API to become stable, i.e., drop `x-` - Wait for qemu to offer notification for migration state change v4: - Rename low-level API flag to `VIR_MIGRATE_ENABLE_POSTCOPY` - Added high-level API flag `VIR_MIGRATE_POSTCOPY_AFTER_PRECOPY` - Do not introduce a new job type and use migration status instead - Added both low- and high-level interface to virsh - Tested with OpenStack Icehouse Cristian Klein (11): Added public API for post-copy migration qemu: added low-level post-copy migration functions qemu: implemented VIR_MIGRATE_ENABLE_POSTCOPY qemu: implemented post-copy migration logic qemu: implement virDomainMigrateStartPostCopy virsh: added --enable-postcopy and migrate-start-postcopy virsh: added --postcopy-after to migrate command qemu: retrieve dirty sync count qemu: implemented VIR_MIGRATE_POSTCOPY_AFTER_PRECOPY virsh: added --postcopy-after-precopy to migrate Revert "Do not allow changing the UUID of a nwfilter" include/libvirt/libvirt-domain.h | 5 ++ src/conf/nwfilter_conf.c | 11 --- src/driver-hypervisor.h | 5 ++ src/libvirt-domain.c | 92 +++++++++++++++++++++ src/libvirt_public.syms | 1 + src/qemu/qemu_driver.c | 60 ++++++++++++++ src/qemu/qemu_migration.c | 169 +++++++++++++++++++++++++++++++++++++-- src/qemu/qemu_migration.h | 4 +- src/qemu/qemu_monitor.c | 24 +++++- src/qemu/qemu_monitor.h | 5 ++ src/qemu/qemu_monitor_json.c | 27 ++++++- src/qemu/qemu_monitor_json.h | 1 + src/qemu/qemu_monitor_text.c | 1 + src/remote/remote_driver.c | 1 + src/remote/remote_protocol.x | 12 ++- src/remote_protocol-structs | 5 ++ tests/qemumonitorjsontest.c | 1 + tools/virsh-domain.c | 116 ++++++++++++++++++++++++++- tools/virsh.pod | 21 +++++ 19 files changed, 536 insertions(+), 25 deletions(-) -- 1.9.1

5 17

[libvirt] [PATCH] Report original error when QMP probing fails with new QEMU
by Daniel P. Berrange 04 Dec '14

04 Dec '14

If probing capabilities via QMP fails, we now have a check that prevents us falling back to -help parsing. Unfortunately the error message "Failed to probe capabilities for /usr/bin/qemu-kvm: unsupported configuration: QEMU 2.1.2 is too new for help parsing" is proving rather unhelpful to the user. We need to be telling them why QMP failed (the root cause), rather than they can't use -help (the side effect). To do this we should capture stderr during QMP probing, and if -help parsing then sees a new QEMU version, we know that QMP should have worked, and so we can show the messages from stderr. The message thus becomes "Failed to probe capabilities for /usr/bin/qemu-kvm: internal error: QEMU / QMP failed: Could not access KVM kernel module: No such file or directory failed to initialize KVM: No such file or directory" --- src/qemu/qemu_capabilities.c | 37 +++++++++++++++++++++++++++---------- src/qemu/qemu_capabilities.h | 3 ++- src/qemu/qemu_process.c | 4 ++-- tests/qemuhelptest.c | 2 +- 4 files changed, 32 insertions(+), 14 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index d2c046d..f777484 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -1331,7 +1331,8 @@ int virQEMUCapsParseHelpStr(const char *qemu, unsigned int *version, bool *is_kvm, unsigned int *kvm_version, - bool check_yajl) + bool check_yajl, + const char *qemuerr) { unsigned major, minor, micro; const char *p = help; @@ -1392,9 +1393,15 @@ int virQEMUCapsParseHelpStr(const char *qemu, * using QMP probing. */ if (*version >= 1002000) { - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("QEMU %u.%u.%u is too new for help parsing"), - major, minor, micro); + if (qemuerr && *qemuerr) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("QEMU / QMP failed: %s"), + qemuerr); + } else { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("QEMU %u.%u.%u is too new for help parsing"), + major, minor, micro); + } goto cleanup; } @@ -2967,7 +2974,7 @@ virQEMUCapsInitCached(virQEMUCapsPtr qemuCaps, const char *cacheDir) #define QEMU_SYSTEM_PREFIX "qemu-system-" static int -virQEMUCapsInitHelp(virQEMUCapsPtr qemuCaps, uid_t runUid, gid_t runGid) +virQEMUCapsInitHelp(virQEMUCapsPtr qemuCaps, uid_t runUid, gid_t runGid, const char *qemuerr) { virCommandPtr cmd = NULL; bool is_kvm; @@ -2998,7 +3005,8 @@ virQEMUCapsInitHelp(virQEMUCapsPtr qemuCaps, uid_t runUid, gid_t runGid) &qemuCaps->version, &is_kvm, &qemuCaps->kvmVersion, - false) < 0) + false, + qemuerr) < 0) goto cleanup; /* x86_64 and i686 support PCI-multibus on all machine types @@ -3249,7 +3257,8 @@ static int virQEMUCapsInitQMP(virQEMUCapsPtr qemuCaps, const char *libDir, uid_t runUid, - gid_t runGid) + gid_t runGid, + char **qemuerr) { int ret = -1; virCommandPtr cmd = NULL; @@ -3311,13 +3320,16 @@ virQEMUCapsInitQMP(virQEMUCapsPtr qemuCaps, virCommandSetGID(cmd, runGid); virCommandSetUID(cmd, runUid); + virCommandSetErrorBuffer(cmd, qemuerr); + /* Log, but otherwise ignore, non-zero status. */ if (virCommandRun(cmd, &status) < 0) goto cleanup; if (status != 0) { ret = 0; - VIR_DEBUG("QEMU %s exited with status %d", qemuCaps->binary, status); + VIR_DEBUG("QEMU %s exited with status %d: %s", + qemuCaps->binary, status, *qemuerr); goto cleanup; } @@ -3366,6 +3378,8 @@ virQEMUCapsInitQMP(virQEMUCapsPtr qemuCaps, VIR_ERROR(_("Failed to kill process %lld: %s"), (long long) pid, virStrerror(errno, ebuf, sizeof(ebuf))); + + VIR_FREE(*qemuerr); } if (pidfile) { unlink(pidfile); @@ -3406,6 +3420,7 @@ virQEMUCapsPtr virQEMUCapsNewForBinary(const char *binary, virQEMUCapsPtr qemuCaps; struct stat sb; int rv; + char *qemuerr = NULL; if (!(qemuCaps = virQEMUCapsNew())) goto error; @@ -3436,13 +3451,13 @@ virQEMUCapsPtr virQEMUCapsNewForBinary(const char *binary, goto error; if (rv == 0) { - if (virQEMUCapsInitQMP(qemuCaps, libDir, runUid, runGid) < 0) { + if (virQEMUCapsInitQMP(qemuCaps, libDir, runUid, runGid, &qemuerr) < 0) { virQEMUCapsLogProbeFailure(binary); goto error; } if (!qemuCaps->usedQMP && - virQEMUCapsInitHelp(qemuCaps, runUid, runGid) < 0) { + virQEMUCapsInitHelp(qemuCaps, runUid, runGid, qemuerr) < 0) { virQEMUCapsLogProbeFailure(binary); goto error; } @@ -3451,9 +3466,11 @@ virQEMUCapsPtr virQEMUCapsNewForBinary(const char *binary, goto error; } + VIR_FREE(qemuerr); return qemuCaps; error: + VIR_FREE(qemuerr); virObjectUnref(qemuCaps); qemuCaps = NULL; return NULL; diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index c5542d1..f53b8f4 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -308,7 +308,8 @@ int virQEMUCapsParseHelpStr(const char *qemu, unsigned int *version, bool *is_kvm, unsigned int *kvm_version, - bool check_yajl); + bool check_yajl, + const char *qemuerr); /* Only for use by test suite */ int virQEMUCapsParseDeviceStr(virQEMUCapsPtr qemuCaps, const char *str); diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index a14b6f7..68777e8 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -3554,12 +3554,12 @@ qemuProcessReconnect(void *opaque) struct qemuProcessReconnectData *data = opaque; virQEMUDriverPtr driver = data->driver; virDomainObjPtr obj = data->obj; - qemuDomainObjPrivatePtr priv; + qemuDomainObjPrivatePtr priv = NULL; virConnectPtr conn = data->conn; struct qemuDomainJobObj oldjob; int state; int reason; - virQEMUDriverConfigPtr cfg; + virQEMUDriverConfigPtr cfg = NULL; size_t i; int ret; diff --git a/tests/qemuhelptest.c b/tests/qemuhelptest.c index 0c68367..6d9525c 100644 --- a/tests/qemuhelptest.c +++ b/tests/qemuhelptest.c @@ -59,7 +59,7 @@ static int testHelpStrParsing(const void *data) goto cleanup; if (virQEMUCapsParseHelpStr("QEMU", help, flags, - &version, &is_kvm, &kvm_version, false) == -1) { + &version, &is_kvm, &kvm_version, false, NULL) == -1) { if (info->error && virGetLastError()->code == info->error) ret = 0; goto cleanup; -- 2.1.0

2 1

[libvirt] [PATCH] docs: storage: fix rbd pool indentation
by Cole Robinson 04 Dec '14

04 Dec '14

--- Pushed as trivial docs/storage.html.in | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/docs/storage.html.in b/docs/storage.html.in index 9933548..f920dae 100644 --- a/docs/storage.html.in +++ b/docs/storage.html.in @@ -551,12 +551,12 @@ <name>myrbdpool</name> <source> <name>rbdpool</name> - <host name='1.2.3.4' port='6789'/> - <host name='my.ceph.monitor' port='6789'/> - <host name='third.ceph.monitor' port='6789'/> - <auth username='admin' type='ceph'> - <secret uuid='2ec115d7-3a88-3ceb-bc12-0ac909a6fd87'/> - </auth> + <host name='1.2.3.4' port='6789'/> + <host name='my.ceph.monitor' port='6789'/> + <host name='third.ceph.monitor' port='6789'/> + <auth username='admin' type='ceph'> + <secret uuid='2ec115d7-3a88-3ceb-bc12-0ac909a6fd87'/> + </auth> </source> </pool></pre> -- 2.1.0

1 0

[libvirt] [PATCH] qemu: process: Avoid uninitialized use two vars when reconnecting to vm
by Peter Krempa 04 Dec '14

04 Dec '14

3ecebf07110ca8d3413072557f29137943e848e3 breaks the build as it adds a way to jump to cleanup before the 'cfg' object is retrieved and 'priv' is initialized. --- Notes: Pushed under the build-breaker rule. src/qemu/qemu_process.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 310a08d..a14b6f7 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -3571,16 +3571,16 @@ qemuProcessReconnect(void *opaque) * deleted if qemuConnectMonitor() failed */ virObjectRef(obj); + cfg = virQEMUDriverGetConfig(driver); + priv = obj->privateData; + if (qemuDomainObjBeginJob(driver, obj, QEMU_JOB_MODIFY) < 0) goto killvm; virNWFilterReadLockFilterUpdates(); - cfg = virQEMUDriverGetConfig(driver); VIR_DEBUG("Reconnect monitor to %p '%s'", obj, obj->def->name); - priv = obj->privateData; - /* XXX check PID liveliness & EXE path */ if (qemuConnectMonitor(driver, obj, QEMU_ASYNC_JOB_NONE, -1) < 0) goto error; -- 2.1.0

1 0

[libvirt] [PATCHv2] qemu: process: Refactor reconnecting to qemu processes
by Peter Krempa 04 Dec '14

04 Dec '14

Move entering the job into the thread to simplify the program flow. Also as the code holds a separate reference to the domain object some conditions can be simplified. --- Notes: Version 2: - fix leak of 'data' when not reconnecting src/qemu/qemu_process.c | 161 ++++++++++++++++++++++-------------------------- 1 file changed, 72 insertions(+), 89 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 08d6b7c..310a08d 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -3524,8 +3524,7 @@ qemuProcessUpdateDevices(virQEMUDriverPtr driver, struct qemuProcessReconnectData { virConnectPtr conn; virQEMUDriverPtr driver; - void *payload; - struct qemuDomainJobObj oldjob; + virDomainObjPtr obj; }; /* * Open an existing VM's monitor, re-detect VCPU threads @@ -3534,13 +3533,27 @@ struct qemuProcessReconnectData { * We own the virConnectPtr we are passed here - whoever started * this thread function has increased the reference counter to it * so that we now have to close it. + * + * This function also inherits a locked domain object. + * + * This function needs to: + * 1. Enter job + * 1. just before monitor reconnect do lightweight MonitorEnter + * (increase VM refcount and unlock VM) + * 2. reconnect to monitor + * 3. do lightweight MonitorExit (lock VM) + * 4. continue reconnect process + * 5. EndJob + * + * We can't do normal MonitorEnter & MonitorExit because these two lock the + * monitor lock, which does not exists in this early phase. */ static void qemuProcessReconnect(void *opaque) { struct qemuProcessReconnectData *data = opaque; virQEMUDriverPtr driver = data->driver; - virDomainObjPtr obj = data->payload; + virDomainObjPtr obj = data->obj; qemuDomainObjPrivatePtr priv; virConnectPtr conn = data->conn; struct qemuDomainJobObj oldjob; @@ -3550,26 +3563,24 @@ qemuProcessReconnect(void *opaque) size_t i; int ret; - memcpy(&oldjob, &data->oldjob, sizeof(oldjob)); - VIR_FREE(data); - virNWFilterReadLockFilterUpdates(); + qemuDomainObjRestoreJob(obj, &oldjob); - virObjectLock(obj); + /* Hold an extra reference because we can't allow 'vm' to be + * deleted if qemuConnectMonitor() failed */ + virObjectRef(obj); + + if (qemuDomainObjBeginJob(driver, obj, QEMU_JOB_MODIFY) < 0) + goto killvm; + + virNWFilterReadLockFilterUpdates(); cfg = virQEMUDriverGetConfig(driver); VIR_DEBUG("Reconnect monitor to %p '%s'", obj, obj->def->name); priv = obj->privateData; - /* Job was started by the caller for us */ - qemuDomainObjTransferJob(obj); - - /* Hold an extra reference because we can't allow 'vm' to be - * deleted if qemuConnectMonitor() failed */ - virObjectRef(obj); - /* XXX check PID liveliness & EXE path */ if (qemuConnectMonitor(driver, obj, QEMU_ASYNC_JOB_NONE, -1) < 0) goto error; @@ -3701,10 +3712,10 @@ qemuProcessReconnect(void *opaque) driver->inhibitCallback(true, driver->inhibitOpaque); endjob: - if (!qemuDomainObjEndJob(driver, obj)) - obj = NULL; + /* we hold an extra reference, so this will never fail */ + ignore_value(qemuDomainObjEndJob(driver, obj)); - if (obj && virObjectUnref(obj)) + if (virObjectUnref(obj)) virObjectUnlock(obj); virObjectUnref(conn); @@ -3714,35 +3725,35 @@ qemuProcessReconnect(void *opaque) return; error: - if (!qemuDomainObjEndJob(driver, obj)) - obj = NULL; - - if (obj) { - if (!virDomainObjIsActive(obj)) { - if (virObjectUnref(obj)) - virObjectUnlock(obj); - } else if (virObjectUnref(obj)) { - /* We can't get the monitor back, so must kill the VM - * to remove danger of it ending up running twice if - * user tries to start it again later - */ - if (virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_NO_SHUTDOWN)) { - /* If we couldn't get the monitor and qemu supports - * no-shutdown, we can safely say that the domain - * crashed ... */ - state = VIR_DOMAIN_SHUTOFF_CRASHED; - } else { - /* ... but if it doesn't we can't say what the state - * really is and FAILED means "failed to start" */ - state = VIR_DOMAIN_SHUTOFF_UNKNOWN; - } - qemuProcessStop(driver, obj, state, 0); - if (!obj->persistent) - qemuDomainRemoveInactive(driver, obj); - else - virObjectUnlock(obj); + /* we hold an extra reference, so this will never fail */ + ignore_value(qemuDomainObjEndJob(driver, obj)); + + killvm: + if (virDomainObjIsActive(obj)) { + /* We can't get the monitor back, so must kill the VM + * to remove danger of it ending up running twice if + * user tries to start it again later + */ + if (virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_NO_SHUTDOWN)) { + /* If we couldn't get the monitor and qemu supports + * no-shutdown, we can safely say that the domain + * crashed ... */ + state = VIR_DOMAIN_SHUTOFF_CRASHED; + } else { + /* ... but if it doesn't we can't say what the state + * really is and FAILED means "failed to start" */ + state = VIR_DOMAIN_SHUTOFF_UNKNOWN; } + qemuProcessStop(driver, obj, state, 0); } + + if (virObjectUnref(obj)) { + if (!obj->persistent) + qemuDomainRemoveInactive(driver, obj); + else + virObjectUnlock(obj); + } + virObjectUnref(conn); virObjectUnref(cfg); virNWFilterUnlockFilterUpdates(); @@ -3756,6 +3767,7 @@ qemuProcessReconnectHelper(virDomainObjPtr obj, struct qemuProcessReconnectData *src = opaque; struct qemuProcessReconnectData *data; + /* If the VM was inactive, we don't need to reconnect */ if (!obj->pid) return 0; @@ -3763,64 +3775,35 @@ qemuProcessReconnectHelper(virDomainObjPtr obj, return -1; memcpy(data, src, sizeof(*data)); - data->payload = obj; - - /* - * We create a separate thread to run qemuProcessReconnect in it. - * However, qemuProcessReconnect needs to: - * 1. just before monitor reconnect do lightweight MonitorEnter - * (increase VM refcount, unlock VM & driver) - * 2. reconnect to monitor - * 3. do lightweight MonitorExit (lock VM) - * 4. continue reconnect process - * 5. EndJob - * - * NB, we can't do normal MonitorEnter & MonitorExit because - * these two lock the monitor lock, which does not exists in - * this early phase. - */ + data->obj = obj; + /* this lock will be eventually transferred to the thread that handles the + * reconnect */ virObjectLock(obj); - qemuDomainObjRestoreJob(obj, &data->oldjob); - - if (qemuDomainObjBeginJob(src->driver, obj, QEMU_JOB_MODIFY) < 0) - goto error; - - /* Since we close the connection later on, we have to make sure - * that the threads we start see a valid connection throughout their - * lifetime. We simply increase the reference counter here. + /* Since we close the connection later on, we have to make sure that the + * threads we start see a valid connection throughout their lifetime. We + * simply increase the reference counter here. */ virObjectRef(data->conn); if (virThreadCreate(&thread, false, qemuProcessReconnect, data) < 0) { - - virObjectUnref(data->conn); - virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Could not create thread. QEMU initialization " "might be incomplete")); - if (!qemuDomainObjEndJob(src->driver, obj)) { - obj = NULL; - } else if (virObjectUnref(obj)) { - /* We can't spawn a thread and thus connect to monitor. - * Kill qemu */ - qemuProcessStop(src->driver, obj, VIR_DOMAIN_SHUTOFF_FAILED, 0); - if (!obj->persistent) - qemuDomainRemoveInactive(src->driver, obj); - else - virObjectUnlock(obj); - } - goto error; - } + /* We can't spawn a thread and thus connect to monitor. Kill qemu. */ + qemuProcessStop(src->driver, obj, VIR_DOMAIN_SHUTOFF_FAILED, 0); + if (!obj->persistent) + qemuDomainRemoveInactive(src->driver, obj); + else + virObjectUnlock(obj); - virObjectUnlock(obj); + virObjectUnref(data->conn); + VIR_FREE(data); + return -1; + } return 0; - - error: - VIR_FREE(data); - return -1; } /** -- 2.1.0

2 2

[libvirt] [PATCH] spec: Remove qemu-img dependency from storage driver
by Jiri Denemark 04 Dec '14

04 Dec '14

Both qemu-img and its stripped down version qcow-create are searched for dynamically with virFindFileInPath, functions using them report errors when the required binary is not present, and we don't compile in a fixed path detected by configure anymore. So I don't see a reason for keeping the dependencies for our storage driver. Moreover, it's weired when the dependencies (of the storage driver) change depending on wheter a particular hypervisor driver (xen/qemu) is enabled or not. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- libvirt.spec.in | 18 ------------------ 1 file changed, 18 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index bda28e7..dbeb62a 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -530,15 +530,6 @@ BuildRequires: PolicyKit-devel >= 0.6 # For mount/umount in FS driver BuildRequires: util-linux %endif -%if %{with_qemu} -# From QEMU RPMs -BuildRequires: /usr/bin/qemu-img -%else - %if %{with_xen} -# From Xen RPMs -BuildRequires: /usr/sbin/qcow-create - %endif -%endif %if %{with_storage_lvm} # For LVM drivers BuildRequires: lvm2 @@ -871,15 +862,6 @@ Requires: device-mapper # For Sheepdog support Requires: sheepdog %endif - %if %{with_qemu} -# From QEMU RPMs -Requires: /usr/bin/qemu-img - %else - %if %{with_xen} -# From Xen RPMs -Requires: /usr/sbin/qcow-create - %endif - %endif %description daemon-driver-storage The storage driver plugin for the libvirtd daemon, providing -- 2.2.0

3 4

[libvirt] [PATCH] drvbhyve: Automatically tear down guest domains on shutdown
by Conrad Meyer 04 Dec '14

04 Dec '14

Reboot requires more sophistication and is left as a future work item -- but at least part of the plumbing is in place. --- Looking for feedback. I'm pretty unfamiliar with libvirt; maybe this isn't quite the right way to do this. Sorry. If you want me to rework it in some way, just let me know. I tried to model this off of DrvQEMU, which I knew to support this behavior. Reboot should probably be processed on a threadpool thread outside of the event loop, so I omitted it for now. P.S., the 'read-non-seekable' check test seems to just hang forever on FreeBSD. I haven't dug into it, but POSIX FIFO behavior is pretty weird. --- src/Makefile.am | 2 + src/bhyve/bhyve_monitor.c | 184 ++++++++++++++++++++++++++++++++++++++++++++++ src/bhyve/bhyve_monitor.h | 36 +++++++++ src/bhyve/bhyve_process.c | 14 +++- 4 files changed, 233 insertions(+), 3 deletions(-) create mode 100644 src/bhyve/bhyve_monitor.c create mode 100644 src/bhyve/bhyve_monitor.h diff --git a/src/Makefile.am b/src/Makefile.am index d8fe624..b6c1701 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -833,6 +833,8 @@ BHYVE_DRIVER_SOURCES = \ bhyve/bhyve_domain.h \ bhyve/bhyve_driver.h \ bhyve/bhyve_driver.c \ + bhyve/bhyve_monitor.c \ + bhyve/bhyve_monitor.h \ bhyve/bhyve_process.c \ bhyve/bhyve_process.h \ bhyve/bhyve_utils.h \ diff --git a/src/bhyve/bhyve_monitor.c b/src/bhyve/bhyve_monitor.c new file mode 100644 index 0000000..cd3cf6e --- /dev/null +++ b/src/bhyve/bhyve_monitor.c @@ -0,0 +1,184 @@ +/* + * bhyve_monitor.c: Tear-down or reboot bhyve domains on guest shutdown + * + * Copyright (C) 2014 Conrad Meyer + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + * + * Author: Conrad Meyer <cse.cem(a)gmail.com> + */ + +#include <config.h> + +#include <sys/types.h> +#include <sys/event.h> +#include <sys/time.h> +#include <sys/wait.h> + +#include "bhyve_monitor.h" +#include "bhyve_process.h" +#include "viralloc.h" +#include "virerror.h" +#include "virfile.h" +#include "virlog.h" + +#define VIR_FROM_THIS VIR_FROM_BHYVE + +VIR_LOG_INIT("bhyve.bhyve_monitor"); + +struct _bhyveMonitor { + int kq; + int watch; + virDomainObjPtr vm; + bhyveConnPtr driver; +}; + +static void +bhyveMonitorIO(int watch, int kq, int events ATTRIBUTE_UNUSED, void *opaque) +{ + const struct timespec zerowait = {}; + bhyveMonitorPtr mon = opaque; + struct kevent kev; + int rc, status; + + if (watch != mon->watch || kq != mon->kq) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("event from unexpected fd %d!=%d / watch %d!=%d"), + mon->kq, kq, mon->watch, watch); + return; + } + + rc = kevent(kq, NULL, 0, &kev, 1, &zerowait); + if (rc < 0) { + virReportSystemError(errno, "%s", _("Unable to query kqueue")); + return; + } + + if (rc == 0) + return; + + if ((kev.flags & EV_ERROR) != 0) { + virReportSystemError(kev.data, "%s", _("Unable to query kqueue")); + return; + } + + if (kev.filter == EVFILT_PROC && (kev.fflags & NOTE_EXIT) != 0) { + if ((pid_t)kev.ident != mon->vm->pid) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("event from unexpected proc %ju!=%ju"), + (uintmax_t)mon->vm->pid, (uintmax_t)kev.ident); + return; + } + + status = kev.data; + if (WIFSIGNALED(status) && WCOREDUMP(status)) { + VIR_ERROR("Guest %s got signal %d and crashed", mon->vm->def->name, + WTERMSIG(status)); + virBhyveProcessStop(mon->driver, mon->vm, + VIR_DOMAIN_SHUTOFF_CRASHED); + } else if (WIFEXITED(status)) { + if (WEXITSTATUS(status) == 0) { + /* 0 - reboot */ + /* TODO: Implementing reboot is a little more complicated. */ + VIR_INFO("Guest %s rebooted; destroying domain.", + mon->vm->def->name); + virBhyveProcessStop(mon->driver, mon->vm, + VIR_DOMAIN_SHUTOFF_SHUTDOWN); + } else if (WEXITSTATUS(status) < 3) { + /* 1 - shutdown, 2 - halt, 3 - triple fault. others - error */ + VIR_INFO("Guest %s shut itself down; destroying domain.", + mon->vm->def->name); + virBhyveProcessStop(mon->driver, mon->vm, + VIR_DOMAIN_SHUTOFF_SHUTDOWN); + } else { + VIR_INFO("Guest %s had an error and exited with status %d; destroying domain.", + mon->vm->def->name, WEXITSTATUS(status)); + virBhyveProcessStop(mon->driver, mon->vm, + VIR_DOMAIN_SHUTOFF_UNKNOWN); + } + } + } +} + +static void +bhyveMonitorRelease(void *opaque) +{ + bhyveMonitorPtr mon = opaque; + + VIR_FORCE_CLOSE(mon->kq); + virObjectUnref(mon->vm); + VIR_FREE(mon); +} + +bhyveMonitorPtr +bhyveMonitorOpen(virDomainObjPtr vm, bhyveConnPtr driver) +{ + bhyveMonitorPtr mon; + struct kevent kev; + int rc; + + if (VIR_ALLOC(mon) < 0) + return NULL; + + mon->vm = virObjectRef(vm); + mon->driver = driver; + + mon->kq = kqueue(); + if (mon->kq < 0) { + virReportError(VIR_ERR_SYSTEM_ERROR, "%s", + _("Unable to create kqueue")); + goto cleanup; + } + + EV_SET(&kev, vm->pid, EVFILT_PROC, EV_ADD, NOTE_EXIT, 0, mon); + rc = kevent(mon->kq, &kev, 1, NULL, 0, NULL); + if (rc < 0) { + virReportError(VIR_ERR_SYSTEM_ERROR, "%s", + _("Unable to register process kevent")); + goto cleanup; + } + + mon->watch = virEventAddHandle(mon->kq, + VIR_EVENT_HANDLE_READABLE | + VIR_EVENT_HANDLE_ERROR | + VIR_EVENT_HANDLE_HANGUP, + bhyveMonitorIO, + mon, + bhyveMonitorRelease); + if (mon->watch < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("unable to register monitor events")); + goto cleanup; + } + + return mon; + + cleanup: + bhyveMonitorRelease(mon); + return NULL; +} + +void +bhyveMonitorClose(bhyveMonitorPtr mon) +{ + + if (mon == NULL) + return; + + if (mon->watch > 0) + virEventRemoveHandle(mon->watch); + else + bhyveMonitorRelease(mon); +} diff --git a/src/bhyve/bhyve_monitor.h b/src/bhyve/bhyve_monitor.h new file mode 100644 index 0000000..226d878 --- /dev/null +++ b/src/bhyve/bhyve_monitor.h @@ -0,0 +1,36 @@ +/* + * bhyve_monitor.h: Tear-down or reboot bhyve domains on guest shutdown + * + * Copyright (C) 2014 Conrad Meyer + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + * + * Author: Conrad Meyer <cse.cem(a)gmail.com> + */ + +#ifndef BHYVE_MONITOR_H +# define BHYVE_MONITOR_H + +# include "internal.h" +# include "domain_conf.h" +# include "bhyve_utils.h" + +typedef struct _bhyveMonitor bhyveMonitor; +typedef bhyveMonitor *bhyveMonitorPtr; + +bhyveMonitorPtr bhyveMonitorOpen(virDomainObjPtr vm, bhyveConnPtr driver); +void bhyveMonitorClose(bhyveMonitorPtr mon); + +#endif /* BHYVE_MONITOR_H */ diff --git a/src/bhyve/bhyve_process.c b/src/bhyve/bhyve_process.c index a30e36a..284641a 100644 --- a/src/bhyve/bhyve_process.c +++ b/src/bhyve/bhyve_process.c @@ -32,8 +32,9 @@ #include <net/if_tap.h> #include "bhyve_device.h" -#include "bhyve_process.h" #include "bhyve_command.h" +#include "bhyve_monitor.h" +#include "bhyve_process.h" #include "datatypes.h" #include "virerror.h" #include "virlog.h" @@ -209,6 +210,7 @@ virBhyveProcessStart(virConnectPtr conn, vm->def->id = vm->pid; virDomainObjSetState(vm, VIR_DOMAIN_RUNNING, reason); + vm->privateData = bhyveMonitorOpen(vm, driver); if (virDomainSaveStatus(driver->xmlopt, BHYVE_STATE_DIR, @@ -268,6 +270,9 @@ virBhyveProcessStop(bhyveConnPtr driver, return -1; } + if (vm->privateData != NULL) + bhyveMonitorClose((bhyveMonitorPtr)vm->privateData); + /* First, try to kill 'bhyve' process */ if (virProcessKillPainfully(vm->pid, true) != 0) VIR_WARN("Failed to gracefully stop bhyve VM '%s' (pid: %d)", @@ -371,9 +376,12 @@ virBhyveProcessReconnect(virDomainObjPtr vm, goto cleanup; proc_argv = kvm_getargv(data->kd, kp, 0); - if (proc_argv && proc_argv[0]) - if (STREQ(expected_proctitle, proc_argv[0])) + if (proc_argv && proc_argv[0]) { + if (STREQ(expected_proctitle, proc_argv[0])) { ret = 0; + vm->privateData = bhyveMonitorOpen(vm, data->driver); + } + } cleanup: if (ret < 0) { -- 1.9.3

2 2

[libvirt] [PATCH 0/3] qemu: support update graphic device persistently
by Wang Rui 04 Dec '14

04 Dec '14

We can change vnc password by using virDomainUpdateDeviceFlags API with live flag. But it can't be changed with config flag. 1/3: improve the error number when update graphics failed 2/3: refactor the function qemuDomainFindGraphics for the future patch 3/3: support updating vnc/spice auth arguments persistently Wang Rui (3): qemu: report properer error number when change graphics failed qemu: revise qemuDomainFindGraphics to be reused in the future patch qemu: make persistent update of graphics device supported src/conf/domain_conf.c | 3 ++- src/qemu/qemu_driver.c | 38 +++++++++++++++++++++++++++++++++++++- src/qemu/qemu_hotplug.c | 23 +++++++++++------------ src/qemu/qemu_hotplug.h | 3 +++ 4 files changed, 53 insertions(+), 14 deletions(-) -- 1.7.12.4

2 5

[libvirt] [PATCHv2 0/9] Let libvirt manage a bridge's FDB
by Laine Stump 04 Dec '14

04 Dec '14

(Everyone - see the request for opinions/ideas towards the bottom) The idea behind these patches is the following: 1) most virtual machines only have a single MAC address behind each interface, and that MAC address is known by libvirt. 2) If we (i.e. libvirt) manually add an entry to the bridge's forwarding database (fdb) for the MAC address associated with a port on the bridge, we can turn off learning and unicast_flooding for that port. 3) kernels starting with 3.15 (and actually working correctly starting in kernel 3.17) will notice that all of a bridge's ports have flood and learning turned off, and in that case will turn off promiscuous mode on all ports. If all but one of the ports have flood/learning turned off, then promiscuous will be turned off on that port (and left on for all the other ports) 4) When (4) can be done, there is a measurable performance advantage. It can also *kind of* help security, as it will prevent a guest from doing anything useful if it changes its MAC address (but won't prevent the guest from *sending* packets with a spoofed MAC address). NB: These only work with a fixed MAC address, and no vlan tags set in the guest. Support for both of those will be coming. HERE IS THE REQUEST FOR OPINIONS/IDEAS: This V2 of the patchset addresses several issues brought up by jferlan on the original series, and changes the name of the attribute from: promiscLinks='yes|no' to fdb='learningWithFlood|managed' I'm somewhat more happy with this new naming than the previous but still looking for better ideas. It is closer to describing what the new code really does, but "learningWithFlood" seems a bit long and awkward, while I have been told that "fdb" is too short and unrecognizeable (I will point out that 1) "fdb" is the same name used by iproute2's "bridge" command, and 2) another bridge option, "stp" is also a three letter acronym that will only be recognized by those familiar with configuring an L2 bridge device or watching NASCAR on Saturday afternoons (or whenever it's on - not a fan myself :-)) Here is a full list of every idea that either I or someone else has come up with since I started thinking about this: promiscLinks='yes|no' After initially going with this for the v1 of the patchset, I later decided against it, because it doesn't describe what libvirt is doing, but only a *possible* side effect on *some* of the ports connected to the bridge (in practice, it only happens to the physdev port). fdb='auto|managed' I like "fdb" as the name of the attribute, because I think it really gets at what libvirt is doing - it is taking over management of the bridge's fdb (Forwarding Database), which ends up providing better performance in several ways. The problem with this proposal is that the two values are kind of ambiguous - it's not clear which one is using the bridge module's built-in management (I had figured this would be "auto"), and which is telling libvirt to manage it ("managed"). (on the other hand, the first option is "ignore the issue, let the underlying system handle it", vs. "libvirt should manage it", so maybe it *is* a reasonable choice). Also, see the comment above about the perceived terseness and obscurity of "fdb". fdb='learningWithFlood|managed' This alternate name was suggested by Michael Tsirkin as a way to unambiguously indicate what was being done in the mode where libvirt isn't involved in the fdb management. There was some criticism on IRC that the name is *too* verbose, especially when contrasted with "fdb". fdb='learning|managed' A suggested shortening of "learningWithFlood". forwardingDatabase='blah' A way to get around criticism of "fdb". I think this is too verbose, but maybe I'm biased :-) [specify each minor item that separately] In order to manage the fdb by itself, libvirt disables "learning" and "unicast_flood" for each tap device, enables "vlan_filtering" for the bridge itself, then adds an fdb entry for each tap to the bridge. There was one suggestion that, rather than trying to come up with a single option that says "do all of these things", we should instead make each of them separately configured. The problem with this is that it makes it too easy to screw up the configuration such that it causes sub-par performance, or simply doesn't work at all. Part of libvirt's job is making it difficult to screw up (or at least easier to succeed); for example, libvirt's virtual networks do a lot of things automatically - create a bridge, add iptables rules for filtering and NAT, run an instance of dnsmasq - over time we've offered the option of tweaking more and more of the details of this setup, but the initial aim was to provide something that worked with as few required (or even permitted) tweaks as possible. I guess what I'm getting at is that I think it would be a mistake to require turning on several different knobs (which individually make little/no sense) in order to get the bridge into this higher performing mode. So - does anyone have an opinion of any of the options offered above, or any ideas for alternates? In the meantime, note that while the default is currently "learningWithFlood" (meaning that that name is never actually directly used/required anywhere, but is just in the RNG and the enum definition), the intent of the people who developed this functionality in the kernel is that eventually it will work so well that libvirt management of the fdb can silently become the default with no visible change in behavior. NOTE: If you want to actually try out these patches, you'll need to apply the following patch which I haven't yet pushed: https://www.redhat.com/archives/libvir-list/2014-November/msg00948.html Also, while the description of V1 stated that patches 08 and 09 were not intended to be pushed yet, due to a problem they caused when restarting libvirtd after an update, that problem has been solved, so I now intend to push patches 08 and 09 along with the rest. Laine Stump (9): util: new functions for setting bridge and bridge port attributes util: functions to manage bridge fdb (forwarding database) conf: new network bridge device attribute fdb network: save bridge name in ActualNetDef when actualType==network too network: store network fdb setting in NetDef actual object network: setup bridge devices for fdb='managed' qemu: setup tap devices for fdb='managed' qemu: always use virDomainNetGetActualBridgeName to get interface's bridge lxc: always use virDomainNetGetActualBridgeName to get interface's bridge docs/formatnetwork.html.in | 42 ++- docs/schemas/network.rng | 9 + src/conf/domain_conf.c | 129 ++++--- src/conf/domain_conf.h | 2 + src/conf/network_conf.c | 50 ++- src/conf/network_conf.h | 11 + src/libvirt_private.syms | 11 + src/lxc/lxc_driver.c | 32 +- src/lxc/lxc_process.c | 32 +- src/network/bridge_driver.c | 74 ++++ src/qemu/qemu_command.c | 53 ++- src/qemu/qemu_hotplug.c | 60 +--- src/util/virnetdevbridge.c | 382 ++++++++++++++++++++- src/util/virnetdevbridge.h | 44 ++- tests/networkxml2xmlin/host-bridge-no-flood.xml | 6 + .../nat-network-explicit-flood.xml | 21 ++ tests/networkxml2xmlout/host-bridge-no-flood.xml | 6 + .../nat-network-explicit-flood.xml | 23 ++ tests/networkxml2xmltest.c | 2 + 19 files changed, 778 insertions(+), 211 deletions(-) create mode 100644 tests/networkxml2xmlin/host-bridge-no-flood.xml create mode 100644 tests/networkxml2xmlin/nat-network-explicit-flood.xml create mode 100644 tests/networkxml2xmlout/host-bridge-no-flood.xml create mode 100644 tests/networkxml2xmlout/nat-network-explicit-flood.xml -- 1.9.3

3 24

[libvirt] [PATCH] Fix handling of whitespae in preprocessor macros for API generator
by Daniel P. Berrange 04 Dec '14

04 Dec '14

The apibuild.py script did not handle whitespace in preprocessor macros, so it failed to detect constants declared with '# define' instead of '#define'. Since we now correctly indent our public header files, we have silently lost all constants from libvirt-api.xml. This also caused us to not detect formatting errors in constant docs --- docs/apibuild.py | 8 ++++++++ include/libvirt/libvirt-host.h | 12 ++++++++---- 2 files changed, 16 insertions(+), 4 deletions(-) diff --git a/docs/apibuild.py b/docs/apibuild.py index 7549a63..9fa9361 100755 --- a/docs/apibuild.py +++ b/docs/apibuild.py @@ -459,6 +459,14 @@ class CLexer: if line[0] == '#': self.tokens = map((lambda x: ('preproc', x)), string.split(line)) + + # We might have whitespace between the '#' and preproc + # macro name, so instead of having a single token element + # of '#define' we might end up with '#' and 'define'. This + # merges them back together + if self.tokens[0][1] == "#": + self.tokens[0] = ('preproc', self.tokens[0][1] + self.tokens[1][1]) + self.tokens = self.tokens[:1] + self.tokens[2:] break l = len(line) if line[0] == '"' or line[0] == "'": diff --git a/include/libvirt/libvirt-host.h b/include/libvirt/libvirt-host.h index 5bd9563..53b529f 100644 --- a/include/libvirt/libvirt-host.h +++ b/include/libvirt/libvirt-host.h @@ -536,28 +536,32 @@ struct _virNodeMemoryStats { */ # define VIR_NODE_MEMORY_SHARED_PAGES_SHARING "shm_pages_sharing" -/* VIR_NODE_MEMORY_SHARED_PAGES_UNSHARED: +/* + * VIR_NODE_MEMORY_SHARED_PAGES_UNSHARED: * * Macro for typed parameter that represents how many pages unique * but repeatedly checked for merging. */ # define VIR_NODE_MEMORY_SHARED_PAGES_UNSHARED "shm_pages_unshared" -/* VIR_NODE_MEMORY_SHARED_PAGES_VOLATILE: +/* + * VIR_NODE_MEMORY_SHARED_PAGES_VOLATILE: * * Macro for typed parameter that represents how many pages changing * too fast to be placed in a tree. */ # define VIR_NODE_MEMORY_SHARED_PAGES_VOLATILE "shm_pages_volatile" -/* VIR_NODE_MEMORY_SHARED_FULL_SCAN: +/* + * VIR_NODE_MEMORY_SHARED_FULL_SCANS: * * Macro for typed parameter that represents how many times all * mergeable areas have been scanned. */ # define VIR_NODE_MEMORY_SHARED_FULL_SCANS "shm_full_scans" -/* VIR_NODE_MEMORY_SHARED_MERGE_ACROSS_NODES: +/* + * VIR_NODE_MEMORY_SHARED_MERGE_ACROSS_NODES: * * Macro for typed parameter that represents whether pages from * different NUMA nodes can be merged. The parameter has type int, -- 2.1.0

2 1

[libvirt] [PATCH] docs: Create html documentation even if XHTML1 DTD is not available to validate
by Ian Campbell 04 Dec '14

04 Dec '14

On a Debian system lacking the w3c-dtd-xhtml package the build fails with: $ make -C docs/ formatcaps.html make: Entering directory '/local/scratch/ianc/devel/libvirt.git/docs' Generating formatcaps.html.tmp I/O error : Attempt to load network entity http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd formatcaps.html.in:2: warning: failed to load external entity "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" ^ I/O error : Attempt to load network entity http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd ../docs/sitemap.html.in:2: warning: failed to load external entity "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" ^ missing XHTML1 DTD rm formatcaps.html.tmp make: Leaving directory '/local/scratch/ianc/devel/libvirt.git/docs' $ ls docs/formatcaps* docs/formatcaps.html.in https://www.redhat.com/archives/libvir-list/2009-November/msg00413.html suggests that the XHTML1 DTD should not be a hard requirement and the docs should be generated but not validated if it is not available. Therefore when the DTD is not available arrange for the .html.tmp file to be propagated to the .html output. Signed-off-by: Ian Campbell <ian.campbell(a)citrix.com> Cc: Daniel Veillard <veillard(a)redhat.com> --- docs/Makefile.am | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index 5485ee9..c5ba688 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -232,7 +232,7 @@ internals/%.html.tmp: internals/%.html.in subsite.xsl page.xsl sitemap.html.in SGML_CATALOG_FILES='$(XML_CATALOG_FILE)' \ $(XMLLINT) --catalogs --nonet --format --valid $< > $(srcdir)/$@ \ || { rm $(srcdir)/$@ && exit 1; }; \ - else echo "missing XHTML1 DTD" ; fi ; fi + else echo "missing XHTML1 DTD"; cat $< > $(srcdir)/$@ ; fi ; fi %.php.tmp: %.php.in site.xsl page.xsl sitemap.html.in @if [ -x $(XSLTPROC) ] ; then \ @@ -258,7 +258,7 @@ html/index.html: libvirt-api.xml newapi.xsl page.xsl sitemap.html.in > /dev/null ; then \ SGML_CATALOG_FILES='$(XML_CATALOG_FILE)' \ $(XMLLINT) --catalogs --nonet --valid --noout $(srcdir)/html/*.html ; \ - else echo "missing XHTML1 DTD" ; fi ; fi + else echo "missing XHTML1 DTD"; cat $< > $(srcdir)/$@ ; fi ; fi $(addprefix $(srcdir)/,$(devhelphtml)): $(srcdir)/libvirt-api.xml $(devhelpxsl) $(AM_V_GEN)if [ -x $(XSLTPROC) ] ; then \ -- 2.1.1

3 2

[libvirt] [libvirt-perl][PATCH] Fix constants for VIR_DOMAIN_GET_ALL_STATS_*
by Zhe Peng 04 Dec '14

04 Dec '14

Constants of GET_ALL_STATS_* incorrect,this patch fix them in Domain.pm --- lib/Sys/Virt/Domain.pm | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/lib/Sys/Virt/Domain.pm b/lib/Sys/Virt/Domain.pm index 5680afa..e5ce2dd 100644 --- a/lib/Sys/Virt/Domain.pm +++ b/lib/Sys/Virt/Domain.pm @@ -3287,39 +3287,39 @@ bulk domain stats from C<Sys::Virt::get_all_domain_stats>. =over 4 -=item Sys::Virt::GET_ALL_STATS_ACTIVE +=item Sys::Virt::Domain::GET_ALL_STATS_ACTIVE Include stats for active domains -=item Sys::Virt::GET_ALL_STATS_INACTIVE +=item Sys::Virt::Domain::GET_ALL_STATS_INACTIVE Include stats for inactive domains -=item Sys::Virt::GET_ALL_STATS_OTHER +=item Sys::Virt::Domain::GET_ALL_STATS_OTHER Include stats for other domains -=item Sys::Virt::GET_ALL_STATS_PAUSED +=item Sys::Virt::Domain::GET_ALL_STATS_PAUSED Include stats for paused domains -=item Sys::Virt::GET_ALL_STATS_PERSISTENT +=item Sys::Virt::Domain::GET_ALL_STATS_PERSISTENT Include stats for persistent domains -=item Sys::Virt::GET_ALL_STATS_RUNNING +=item Sys::Virt::Domain::GET_ALL_STATS_RUNNING Include stats for running domains -=item Sys::Virt::GET_ALL_STATS_SHUTOFF +=item Sys::Virt::Domain::GET_ALL_STATS_SHUTOFF Include stats for shutoff domains -=item Sys::Virt::GET_ALL_STATS_TRANSIENT +=item Sys::Virt::Domain::GET_ALL_STATS_TRANSIENT Include stats for transient domains -=item Sys::Virt::GET_ALL_STATS_ENFORCE_STATS +=item Sys::Virt::Domain::GET_ALL_STATS_ENFORCE_STATS Require that all requested stats fields are returned -- 1.9.0

2 1

[libvirt] [PATCH] network: don't allow multiple dhcp sections
by Kyle DeFrancia 04 Dec '14

04 Dec '14

This resolves: https://bugzilla.redhat.com/show_bug.cgi?id=907779 A <dhcp> element can exist in only one IPv4 address and one IPv6 address per network. This patch enforces that in virNetworkUpdate. --- src/conf/network_conf.c | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index 067334e..a914962 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -3472,6 +3472,31 @@ virNetworkIpDefByIndex(virNetworkDefPtr def, int parentIndex) } static int +virNetworkDefUpdateCheckMultiDHCP(virNetworkDefPtr def, + virNetworkIpDefPtr ipdef) +{ + int family = VIR_SOCKET_ADDR_FAMILY(&ipdef->address); + size_t i; + virNetworkIpDefPtr ip; + + for (i = 0; + (ip = virNetworkDefGetIpByIndex(def, family, i)); + i++) { + if (ip != ipdef) { + if (ip->nranges || ip->nhosts) { + virReportError(VIR_ERR_OPERATION_INVALID, + _("dhcp is supported only for a " + "single %s address on each network"), + (family == AF_INET) ? "IPv4" : "IPv6"); + return -1; + } + } + } + + return 0; +} + +static int virNetworkDefUpdateIPDHCPHost(virNetworkDefPtr def, unsigned int command, int parentIndex, @@ -3536,6 +3561,9 @@ virNetworkDefUpdateIPDHCPHost(virNetworkDefPtr def, } else if ((command == VIR_NETWORK_UPDATE_COMMAND_ADD_FIRST) || (command == VIR_NETWORK_UPDATE_COMMAND_ADD_LAST)) { + if (virNetworkDefUpdateCheckMultiDHCP(def, ipdef) < 0) + goto cleanup; + /* log error if an entry with same name/address/ip already exists */ for (i = 0; i < ipdef->nhosts; i++) { if ((host.mac && @@ -3643,6 +3671,9 @@ virNetworkDefUpdateIPDHCPRange(virNetworkDefPtr def, if ((command == VIR_NETWORK_UPDATE_COMMAND_ADD_FIRST) || (command == VIR_NETWORK_UPDATE_COMMAND_ADD_LAST)) { + if (virNetworkDefUpdateCheckMultiDHCP(def, ipdef) < 0) + goto cleanup; + if (i < ipdef->nranges) { char *startip = virSocketAddrFormat(&range.start); char *endip = virSocketAddrFormat(&range.end); -- 2.1.0

2 1

[libvirt] [PATCH] cpu-driver: Fix the cross driver function call
by Daniel Hansel 04 Dec '14

04 Dec '14

For Intel and PowerPC the implementation is calling a cpu driver function across driver layers (i.e. from qemu driver directly to cpu driver). The correct behavior is to use libvirt API functionality to perform such a inter-driver call. This patch introduces a new cpu driver API function getModels() to retrieve the cpu models. The currect implementation to process the cpu_map XML content is transferred to the INTEL and PowerPC cpu driver specific API functions. Additionally processing the cpu_map XML file is not safe due to the fact that the cpu map does not exist for all architectures. Therefore it is better to encapsulate the processing in the architecture specific cpu drivers. Signed-off-by: Daniel Hansel <daniel.hansel(a)linux.vnet.ibm.com> Reviewed-by: Boris Fiuczynski <fiuczy(a)linux.vnet.ibm.com> Reviewed-by: Viktor Mihajlovski <mihajlov(a)linux.vnet.ibm.com> --- src/cpu/cpu.c | 68 +++++++++------------------------------------------ src/cpu/cpu.h | 4 +++ src/cpu/cpu_powerpc.c | 37 ++++++++++++++++++++++++++++ src/cpu/cpu_x86.c | 33 +++++++++++++++++++++++++ 4 files changed, 86 insertions(+), 56 deletions(-) diff --git a/src/cpu/cpu.c b/src/cpu/cpu.c index 08bec5e..788f688 100644 --- a/src/cpu/cpu.c +++ b/src/cpu/cpu.c @@ -724,43 +724,6 @@ cpuModelIsAllowed(const char *model, return false; } -struct cpuGetModelsData -{ - char **data; - size_t len; /* It includes the last element of DATA, which is NULL. */ -}; - -static int -cpuGetArchModelsCb(cpuMapElement element, - xmlXPathContextPtr ctxt, - void *cbdata) -{ - char *name; - struct cpuGetModelsData *data = cbdata; - if (element != CPU_MAP_ELEMENT_MODEL) - return 0; - - name = virXPathString("string(@name)", ctxt); - if (name == NULL) - return -1; - - if (!data->data) { - VIR_FREE(name); - data->len++; - return 0; - } - - return VIR_INSERT_ELEMENT(data->data, data->len - 1, data->len, name); -} - - -static int -cpuGetArchModels(const char *arch, struct cpuGetModelsData *data) -{ - return cpuMapLoad(arch, cpuGetArchModelsCb, data); -} - - /** * cpuGetModels: * @@ -774,18 +737,17 @@ cpuGetArchModels(const char *arch, struct cpuGetModelsData *data) int cpuGetModels(const char *archName, char ***models) { - struct cpuGetModelsData data; - virArch arch; struct cpuArchDriver *driver; - data.data = NULL; - data.len = 1; + virArch arch; + + VIR_DEBUG("arch=%s", archName); arch = virArchFromString(archName); if (arch == VIR_ARCH_NONE) { virReportError(VIR_ERR_INVALID_ARG, _("cannot find architecture %s"), archName); - goto error; + return -1; } driver = cpuGetSubDriver(arch); @@ -793,21 +755,15 @@ cpuGetModels(const char *archName, char ***models) virReportError(VIR_ERR_INVALID_ARG, _("cannot find a driver for the architecture %s"), archName); - goto error; + return -1; } - if (models && VIR_ALLOC_N(data.data, data.len) < 0) - goto error; - - if (cpuGetArchModels(driver->name, &data) < 0) - goto error; - - if (models) - *models = data.data; - - return data.len - 1; + if (!driver->getModels) { + virReportError(VIR_ERR_NO_SUPPORT, + _("CPU driver for %s has no CPU model support"), + virArchToString(arch)); + return -1; + } - error: - virStringFreeList(data.data); - return -1; + return driver->getModels(models); } diff --git a/src/cpu/cpu.h b/src/cpu/cpu.h index 339964c..09e9538 100644 --- a/src/cpu/cpu.h +++ b/src/cpu/cpu.h @@ -100,6 +100,9 @@ typedef char * typedef virCPUDataPtr (*cpuArchDataParse) (const char *xmlStr); +typedef int +(*cpuArchGetModels) (char ***models); + struct cpuArchDriver { const char *name; const virArch *arch; @@ -115,6 +118,7 @@ struct cpuArchDriver { cpuArchHasFeature hasFeature; cpuArchDataFormat dataFormat; cpuArchDataParse dataParse; + cpuArchGetModels getModels; }; diff --git a/src/cpu/cpu_powerpc.c b/src/cpu/cpu_powerpc.c index 67cb9ff..155d93e 100644 --- a/src/cpu/cpu_powerpc.c +++ b/src/cpu/cpu_powerpc.c @@ -649,6 +649,42 @@ ppcBaseline(virCPUDefPtr *cpus, goto cleanup; } +static int +ppcGetModels(char ***models) +{ + struct ppc_map *map; + struct ppc_model *model; + char *name; + size_t nmodels = 0; + + if (!(map = ppcLoadMap())) + goto error; + + if (models && VIR_ALLOC_N(*models, 0) < 0) + goto error; + + model = map->models; + while (model != NULL) { + if (VIR_STRDUP(name, model->name) < 0) + goto error; + + if (VIR_INSERT_ELEMENT(*models, 0, nmodels, name) < 0) + goto error; + + model = model->next; + } + + cleanup: + ppcMapFree(map); + + return nmodels; + + error: + virStringFreeList(*models); + nmodels = -1; + goto cleanup; +} + struct cpuArchDriver cpuDriverPowerPC = { .name = "ppc64", .arch = archs, @@ -662,4 +698,5 @@ struct cpuArchDriver cpuDriverPowerPC = { .baseline = ppcBaseline, .update = ppcUpdate, .hasFeature = NULL, + .getModels = ppcGetModels, }; diff --git a/src/cpu/cpu_x86.c b/src/cpu/cpu_x86.c index 026b54e..f6dcba4 100644 --- a/src/cpu/cpu_x86.c +++ b/src/cpu/cpu_x86.c @@ -2160,6 +2160,38 @@ x86HasFeature(const virCPUData *data, return ret; } +static int +x86GetModels(char ***models) +{ + const struct x86_map *map; + struct x86_model *model; + char *name; + size_t nmodels = 0; + + if (!(map = virCPUx86GetMap())) + return -1; + + if (models && VIR_ALLOC_N(*models, 0) < 0) + goto error; + + model = map->models; + while (model != NULL) { + if (VIR_STRDUP(name, model->name) < 0) + goto error; + + if (VIR_INSERT_ELEMENT(*models, 0, nmodels, name) < 0) + goto error; + + model = model->next; + } + + return nmodels; + + error: + virStringFreeList(*models); + return -1; +} + struct cpuArchDriver cpuDriverX86 = { .name = "x86", @@ -2180,4 +2212,5 @@ struct cpuArchDriver cpuDriverX86 = { .hasFeature = x86HasFeature, .dataFormat = x86CPUDataFormat, .dataParse = x86CPUDataParse, + .getModels = x86GetModels, }; -- 1.8.5.5

4 5

[libvirt] [PATCH] qemu: Fix virsh freeze when blockcopy storage file is removed
by Erik Skultety 04 Dec '14

04 Dec '14

If someone removes blockcopy storage file when still in mirroring phase and then requesting blockjob abort using pivot, virsh cmd freezes. This is not an issue with older qemu versions which did not support asynchronous jobs (which we prefer by default). As we have reached the mirroring phase successfully, polling monitor for blockjob info always returns 1 and the loop never ends. This fix introduces a check for qemuDomainBlockPivot return code, possibly skipping the asynchronous waiting completely, if an error occured and asynchronous waiting was the preferred method. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1139567 --- src/qemu/qemu_driver.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 5dc62b0..2e11938 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -15606,6 +15606,8 @@ qemuDomainBlockJobImpl(virDomainObjPtr vm, if (disk->mirror && (flags & VIR_DOMAIN_BLOCK_JOB_ABORT_PIVOT)) { ret = qemuDomainBlockPivot(conn, driver, vm, device, disk); + if (ret < 0 && async) + goto endjob; goto waitjob; } if (disk->mirror) { -- 1.9.3

3 2

[libvirt] [PATCH] qemu: output error when try to hotplug unsupport console
by Luyao Huang 04 Dec '14

04 Dec '14

https://bugzilla.redhat.com/show_bug.cgi?id=1164627 When try to use attach-device to hotplug a qemu unsupport sonsole, command will return success, and add a console in vm's xml.But this doesn't work in qemu, qemu doesn't support these console.Add a error output when try to hotplug a unsupport console in qemuBuildConsoleChrDeviceStr. Signed-off-by: Luyao Huang <lhuang(a)redhat.com> --- src/qemu/qemu_command.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 1399ce4..2bf4a83 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -10069,13 +10069,17 @@ qemuBuildConsoleChrDeviceStr(char **deviceStr, break; case VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL: + break; case VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_NONE: case VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_XEN: case VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_UML: case VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_LXC: case VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_OPENVZ: case VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_LAST: - break; + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("unsupported console target type %s"), + NULLSTR(virDomainChrConsoleTargetTypeToString(chr->targetType))); + return ret; } ret = 0; -- 1.8.3.1

4 4

[libvirt] [PATCH] network: Let domains be restricted to local DNS
by Josh Stone 04 Dec '14

04 Dec '14

This adds a new "localOnly" attribute on the domain element of the network xml. With this set to "yes", DNS requests under that domain will only be resolved by libvirt's dnsmasq, never forwarded upstream. This was how it worked before commit f69a6b987d616, and I found that functionality useful. For example, I have my host's NetworkManager dnsmasq configured to forward that domain to libvirt's dnsmasq, so I can easily resolve guest names from outside. But if libvirt's dnsmasq doesn't know a name and forwards it to the host, I'd get an endless forwarding loop. Now I can set localOnly="yes" to prevent the loop. Signed-off-by: Josh Stone <jistone(a)redhat.com> Cc: Laine Stump <laine(a)laine.org> --- docs/formatnetwork.html.in | 12 +++++++++++- docs/schemas/network.rng | 3 +++ src/conf/network_conf.c | 5 +++++ src/conf/network_conf.h | 1 + src/network/bridge_driver.c | 5 +++++ .../networkxml2confdata/nat-network-dns-local-domain.conf | 14 ++++++++++++++ tests/networkxml2confdata/nat-network-dns-local-domain.xml | 9 +++++++++ tests/networkxml2conftest.c | 1 + 8 files changed, 49 insertions(+), 1 deletion(-) create mode 100644 tests/networkxml2confdata/nat-network-dns-local-domain.conf create mode 100644 tests/networkxml2confdata/nat-network-dns-local-domain.xml diff --git a/docs/formatnetwork.html.in b/docs/formatnetwork.html.in index dc438aee8622..defcdba00930 100644 --- a/docs/formatnetwork.html.in +++ b/docs/formatnetwork.html.in @@ -82,7 +82,7 @@ <pre> ... <bridge name="virbr0" stp="on" delay="5"/> - <domain name="example.com"/> + <domain name="example.com" localOnly="no"/> <forward mode="nat" dev="eth0"/> ...</pre> @@ -113,6 +113,16 @@ a <code><forward></code> mode of "nat" or "route" (or an isolated network with no <code><forward></code> element). Since 0.4.5 + + + If the optional <code>localOnly</code> attribute on the + <code>domain</code> element is "yes", then DNS requests under + this domain will only be resolved by the virtual network's own + DNS server - they will not be forwarded to the host's upstream + DNS server. If <code>localOnly</code> is "no", and by + default, unresolved requests will be forwarded. + Since 1.2.11 + </dd> <dt><code>forward</code></dt> <dd>Inclusion of the <code>forward</code> element indicates that diff --git a/docs/schemas/network.rng b/docs/schemas/network.rng index 4546f8037580..a1da28092375 100644 --- a/docs/schemas/network.rng +++ b/docs/schemas/network.rng @@ -225,6 +225,9 @@ <optional> <element name="domain"> <attribute name="name"><ref name="dnsName"/></attribute> + <optional> + <attribute name="localOnly"><ref name="virYesNo"/></attribute> + </optional> </element> </optional> diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index 067334e87cb0..61451c39805f 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -2083,6 +2083,11 @@ virNetworkDefParseXML(xmlXPathContextPtr ctxt) /* Parse network domain information */ def->domain = virXPathString("string(./domain[1]/@name)", ctxt); + tmp = virXPathString("string(./domain[1]/@localOnly)", ctxt); + if (tmp) { + def->domain_local = STRCASEEQ(tmp, "yes"); + VIR_FREE(tmp); + } if ((bandwidthNode = virXPathNode("./bandwidth", ctxt)) != NULL && (def->bandwidth = virNetDevBandwidthParse(bandwidthNode, -1)) == NULL) diff --git a/src/conf/network_conf.h b/src/conf/network_conf.h index 660cd2d10cd1..6308a7dcfbf7 100644 --- a/src/conf/network_conf.h +++ b/src/conf/network_conf.h @@ -232,6 +232,7 @@ struct _virNetworkDef { char *bridge; /* Name of bridge device */ char *domain; + bool domain_local; /* Choose not to forward dns for this domain */ unsigned long delay; /* Bridge forward delay (ms) */ bool stp; /* Spanning tree protocol */ virMacAddr mac; /* mac address of bridge device */ diff --git a/src/network/bridge_driver.c b/src/network/bridge_driver.c index 6cb421c52850..dfa375d3aa72 100644 --- a/src/network/bridge_driver.c +++ b/src/network/bridge_driver.c @@ -912,6 +912,11 @@ networkDnsmasqConfContents(virNetworkObjPtr network, } if (network->def->domain) { + if (network->def->domain_local) { + virBufferAsprintf(&configbuf, + "local=/%s/\n", + network->def->domain); + } virBufferAsprintf(&configbuf, "domain=%s\n" "expand-hosts\n", diff --git a/tests/networkxml2confdata/nat-network-dns-local-domain.conf b/tests/networkxml2confdata/nat-network-dns-local-domain.conf new file mode 100644 index 000000000000..5f41b9186cbc --- /dev/null +++ b/tests/networkxml2confdata/nat-network-dns-local-domain.conf @@ -0,0 +1,14 @@ +##WARNING: THIS IS AN AUTO-GENERATED FILE. CHANGES TO IT ARE LIKELY TO BE +##OVERWRITTEN AND LOST. Changes to this configuration should be made using: +## virsh net-edit default +## or other application using the libvirt API. +## +## dnsmasq conf file created by libvirt +strict-order +local=/example.com/ +domain=example.com +expand-hosts +except-interface=lo +bind-dynamic +interface=virbr0 +addn-hosts=/var/lib/libvirt/dnsmasq/default.addnhosts diff --git a/tests/networkxml2confdata/nat-network-dns-local-domain.xml b/tests/networkxml2confdata/nat-network-dns-local-domain.xml new file mode 100644 index 000000000000..a92d71f1f2f6 --- /dev/null +++ b/tests/networkxml2confdata/nat-network-dns-local-domain.xml @@ -0,0 +1,9 @@ +<network> + <name>default</name> + <uuid>81ff0d90-c91e-6742-64da-4a736edb9a9c</uuid> + <forward dev='eth0' mode='nat'/> + <bridge name='virbr0' stp='on' delay='0' /> + <domain name='example.com' localOnly='yes'/> + <ip address='192.168.122.1' netmask='255.255.255.0'> + </ip> +</network> diff --git a/tests/networkxml2conftest.c b/tests/networkxml2conftest.c index 4f1d9345ffe4..d2aa8c62cfcd 100644 --- a/tests/networkxml2conftest.c +++ b/tests/networkxml2conftest.c @@ -146,6 +146,7 @@ mymain(void) DO_TEST("nat-network-dns-hosts", full); DO_TEST("nat-network-dns-forward-plain", full); DO_TEST("nat-network-dns-forwarders", full); + DO_TEST("nat-network-dns-local-domain", full); DO_TEST("dhcp6-network", dhcpv6); DO_TEST("dhcp6-nat-network", dhcpv6); DO_TEST("dhcp6host-routed-network", dhcpv6); -- 2.1.0

4 9

[libvirt] [PATCH] docs: More html/docs changes from libvirt.h.in split
by John Ferlan 04 Dec '14

04 Dec '14

This changes the display from: libvirt-storage: APIs for management of storages to libvirt-storage: APIs for management of storage pools and volumes In making that change I expected my build tree html output to be regenerated; however, it wasn't because the dependency for the separated libvirt-storage.h wasn't there. It was only present for libvirt.h.in So I added each in the order displayed on the docs/html/index.html page Signed-off-by: John Ferlan <jferlan(a)redhat.com> --- docs/Makefile.am | 11 +++++++++++ include/libvirt/libvirt-storage.h | 4 ++-- 2 files changed, 13 insertions(+), 2 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index bb52c35..d624827 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -291,6 +291,17 @@ $(python_generated_files): $(APIBUILD_STAMP) $(APIBUILD_STAMP): $(srcdir)/apibuild.py \ $(srcdir)/../include/libvirt/libvirt.h.in \ + $(srcdir)/../include/libvirt/libvirt-domain-snapshot.h \ + $(srcdir)/../include/libvirt/libvirt-domain.h \ + $(srcdir)/../include/libvirt/libvirt-event.h \ + $(srcdir)/../include/libvirt/libvirt-host.h \ + $(srcdir)/../include/libvirt/libvirt-interface.h \ + $(srcdir)/../include/libvirt/libvirt-network.h \ + $(srcdir)/../include/libvirt/libvirt-nodedev.h \ + $(srcdir)/../include/libvirt/libvirt-nwfilter.h \ + $(srcdir)/../include/libvirt/libvirt-secret.h \ + $(srcdir)/../include/libvirt/libvirt-storage.h \ + $(srcdir)/../include/libvirt/libvirt-stream.h \ $(srcdir)/../include/libvirt/libvirt-lxc.h \ $(srcdir)/../include/libvirt/libvirt-qemu.h \ $(srcdir)/../include/libvirt/virterror.h \ diff --git a/include/libvirt/libvirt-storage.h b/include/libvirt/libvirt-storage.h index 7b2a72d..1f3087b 100644 --- a/include/libvirt/libvirt-storage.h +++ b/include/libvirt/libvirt-storage.h @@ -1,7 +1,7 @@ /* * libvirt-storage.h - * Summary: APIs for management of storages - * Description: Provides APIs for the management of storages + * Summary: APIs for management of storage pools and volumes + * Description: Provides APIs for the management of storage pools and volumes * Author: Daniel Veillard <veillard(a)redhat.com> * * Copyright (C) 2006-2014 Red Hat, Inc. -- 1.9.3

2 1

[libvirt] how to pin disk order when adding SATA disk with virt-manager ?
by Jason Vas Dias 04 Dec '14

04 Dec '14

Hi - this is my first post to this list - sorry if this is a newbie question, but: I want to mount 7 LVM disk volumes under my guest : Physical Host Volume: Guest Device: Guest FS: /dev/mapper/VG_A-vol1 : /dev/sda : /root file system /dev/mapper/VG_S-swap2 : swap : swap partition /dev/mapper/VG_C-vol3 : /dev/sdc : /home file system /dev/mapper/VG_G-vol7 : /dev/sdg : /x some necessary other storage I am booting with the "Direct Boot" option and kernel command line containing "root=/dev/sda" - ie. there HAS to be some particular /dev/sdX where X is the virtual order of the disks on the virtual SATA bus / device minor number to letter mapping . Every time I add a new volume via the virt-manager the ordering of ALL disks change , and I must manually discover which drive letter ('a-g' in /dev/sd[a-g]) maps to which disk , even though the ordering as shown in the GUI remains the same - ie. when I click on the 'SATA Disk 1' icon in the virt-manager 'Details...' tab, it still shows the correct LVM is associated with Disk 1, but the system does not boot because what was '/dev/sda' is now '/dev/sdb' . Is this a bug ? Can I request a "New Feature" which would allow me to specify the "Bus Sequence Number" in the Virtual Disk panel (where I currently can specify Source Path, Disk Bus, Advanced Options etc... ) - surely you could let users map disks to devices in a way that does not change with each new disk added ? Anyway, that's my 2 cents on $subject . Any thoughts / suggestions gratefully received . Thanks & Regards , Jason

1 0

[libvirt] [PATCH] cpu: fix possible crash in getModels
by Pavel Hrdina 04 Dec '14

04 Dec '14

Commit 86a15a25 introduced a new cpu driver API 'getModels'. Public API allow you to pass NULL for models to get only number of existing models. However the new code will crash with segfault so we have to count with the possibility that the user wants only the number. There is also difference in order of the models gathered by this new API as the old approach was inserting the elements to the end of the array so we should use 'VIR_APPEND_ELEMENT'. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/cpu/cpu_powerpc.c | 15 ++++++++++----- src/cpu/cpu_x86.c | 15 ++++++++++----- 2 files changed, 20 insertions(+), 10 deletions(-) diff --git a/src/cpu/cpu_powerpc.c b/src/cpu/cpu_powerpc.c index 871401b..86f80b3 100644 --- a/src/cpu/cpu_powerpc.c +++ b/src/cpu/cpu_powerpc.c @@ -666,11 +666,15 @@ ppcGetModels(char ***models) model = map->models; while (model != NULL) { - if (VIR_STRDUP(name, model->name) < 0) - goto error; + if (models) { + if (VIR_STRDUP(name, model->name) < 0) + goto error; - if (VIR_INSERT_ELEMENT(*models, 0, nmodels, name) < 0) - goto error; + if (VIR_APPEND_ELEMENT(*models, nmodels, name) < 0) + goto error; + } else { + nmodels++; + } model = model->next; } @@ -681,7 +685,8 @@ ppcGetModels(char ***models) return nmodels; error: - virStringFreeList(*models); + if (models) + virStringFreeList(*models); nmodels = -1; goto cleanup; } diff --git a/src/cpu/cpu_x86.c b/src/cpu/cpu_x86.c index f6dcba4..dfbc16c 100644 --- a/src/cpu/cpu_x86.c +++ b/src/cpu/cpu_x86.c @@ -2176,11 +2176,15 @@ x86GetModels(char ***models) model = map->models; while (model != NULL) { - if (VIR_STRDUP(name, model->name) < 0) - goto error; + if (models) { + if (VIR_STRDUP(name, model->name) < 0) + goto error; - if (VIR_INSERT_ELEMENT(*models, 0, nmodels, name) < 0) - goto error; + if (VIR_APPEND_ELEMENT(*models, nmodels, name) < 0) + goto error; + } else { + nmodels++; + } model = model->next; } @@ -2188,7 +2192,8 @@ x86GetModels(char ***models) return nmodels; error: - virStringFreeList(*models); + if (models) + virStringFreeList(*models); return -1; } -- 2.0.4

2 2

[libvirt] [PATCH 0/2] qemu: Refactor reconnecting after daemon restart
by Peter Krempa 04 Dec '14

04 Dec '14

Peter Krempa (2): qemu: driver: Reload snapshots and managedsaves prior to reconnecting qemu: process: Refactor reconnecting to qemu processes src/qemu/qemu_driver.c | 4 +- src/qemu/qemu_process.c | 167 ++++++++++++++++++++++-------------------------- 2 files changed, 78 insertions(+), 93 deletions(-) -- 2.1.0

2 4

[libvirt] [PATCH 0/2] leaseshelper: Coverity/whitespace fixes
by Peter Krempa 04 Dec '14

04 Dec '14

Peter Krempa (2): leaseshelper: Skip entries missing expiry time on INIT action leaseshelper: Fix incorrect alignment of a switch case src/network/leaseshelper.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) -- 2.1.0

2 5

[libvirt] [PATCH] tests: Fix misplaced parenthesis in qemumonitorjsontest
by Martin Kletzander 04 Dec '14

04 Dec '14

When trying clang, it found out that we were comparing sizeof with 0 even though we wanted to check the return value of memcmp. Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- tests/qemumonitorjsontest.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/qemumonitorjsontest.c b/tests/qemumonitorjsontest.c index 5bfcd20..d3ae29a 100644 --- a/tests/qemumonitorjsontest.c +++ b/tests/qemumonitorjsontest.c @@ -1879,7 +1879,7 @@ testQemuMonitorJSONqemuMonitorJSONSetBlockIoThrottle(const void *data) "drive-virtio-disk0", &info, false) < 0) goto cleanup; - if (memcmp(&info, &expectedInfo, sizeof(info) != 0)) { + if (memcmp(&info, &expectedInfo, sizeof(info)) != 0) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", "Invalid @info"); goto cleanup; -- 2.1.3

3 3

[libvirt] [PATCHv5 0/3] leaseshelper: Couple of improvements
by Peter Krempa 04 Dec '14

04 Dec '14

Nehal J Wani (1): leaseshelper: improvements to support all events Peter Krempa (2): leaseshelper: Refactor control flow network: dnsmasq: Don't format lease file path src/network/bridge_driver.c | 16 +- src/network/bridge_driver.h | 3 - src/network/leaseshelper.c | 246 +++++++++++++++------ tests/networkxml2confdata/dhcp6-nat-network.conf | 1 - tests/networkxml2confdata/dhcp6-network.conf | 1 - tests/networkxml2confdata/isolated-network.conf | 1 - .../nat-network-dns-srv-record-minimal.conf | 1 - .../nat-network-dns-srv-record.conf | 1 - .../nat-network-dns-txt-record.conf | 1 - tests/networkxml2confdata/nat-network.conf | 1 - tests/networkxml2confdata/netboot-network.conf | 1 - .../networkxml2confdata/netboot-proxy-network.conf | 1 - tests/networkxml2conftest.c | 12 - 13 files changed, 181 insertions(+), 105 deletions(-) -- 2.1.0

4 8

[libvirt] [PATCH] storage: fix crash caused by no check return before set close
by Luyao Huang 04 Dec '14

04 Dec '14

https://bugzilla.redhat.com/show_bug.cgi?id=1087104#c5 When try to use a invalid offset do volupload, libvirt will get failed in virFDStreamOpenFileInternal, but seems libvirt do not check the return in storageVolUpload, and call virFDStreamSetInternalCloseCb , but stream doesn't have a privateData (is NULL), then crash. Although libvirt have a check for invalid offset in cmdVolUpload, but i think there should have some other way to touch off this crash. 0 0x00007f09429a9c10 in pthread_mutex_lock () from /lib64/libpthread.so.0 1 0x00007f094514dbf5 in virMutexLock (m=<optimized out>) at util/virthread.c:88 2 0x00007f09451cb211 in virFDStreamSetInternalCloseCb at fdstream.c:795 3 0x00007f092ff2c9eb in storageVolUpload at storage/storage_driver.c:2098 4 0x00007f09451f46e0 in virStorageVolUpload at libvirt.c:14000 5 0x00007f0945c78fa1 in remoteDispatchStorageVolUpload at remote_dispatch.h:14339 6 remoteDispatchStorageVolUploadHelper at remote_dispatch.h:14309 7 0x00007f094524a192 in virNetServerProgramDispatchCall at rpc/virnetserverprogram.c:437 Signed-off-by: Luyao Huang <lhuang(a)redhat.com> --- src/storage/storage_driver.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c index 7f33d6f..7f4de19 100644 --- a/src/storage/storage_driver.c +++ b/src/storage/storage_driver.c @@ -2111,8 +2111,9 @@ storageVolUpload(virStorageVolPtr obj, goto cleanup; } - ret = backend->uploadVol(obj->conn, pool, vol, stream, - offset, length, flags); + if ((ret = backend->uploadVol(obj->conn, pool, vol, stream, + offset, length, flags)) < 0) + goto cleanup; /* Add cleanup callback - call after uploadVol since the stream * is then fully set up -- 1.8.3.1

2 1

[libvirt] [PATCH 0/2] Rework reference locking; in QEMU, for starters
by Martin Kletzander 03 Dec '14

03 Dec '14

Our reference locking for objects is great and powerful thing. The problem is that it was either not followed through completely or the design was not complete, but it doesn't matter now. There are few bugs in the code due to the reference counting and the daemon is lacking some performance in specific scenarios. This "series" tried to fix that using the following idea: - Each API working with a domain object that has to get it from the list will have its *own* reference, not borrowed one from the list. - When adding a domain into the list, the reference counter is increased (this is the reference that is there just for being in the list) and when being removed, it is decreased. No special-casing of "if this is was the last reference" and other funny stuff. - When job is created, there is no need to increase the reference counter as there are at least two references for the domain: 1) The API that created the job has one, so if it's not async it will be kept until the API ends and at that point the job won't exist any more. 2) The domain list has one and even though I said nobody needs to rely on that, async APIs probably will do that, but there's an excuse for that. In order to remove the domain from the list, you need a job and that won't succeed unless the async one ended. So we're good in this case as well. After searching through the code for all things that needed to be removed and fixing everything I could possibly think of, I tried a few things on my setup and it looks like it works. However, I haven't tried *every single API*, but I hope that's understandable. On the other hand, I asked Pavel to try running virt-test with these patches applied, hopefully we'll get an idea about how reliable this "series" is. I used "series" (with quotes) on purpose, because first patch just adds two new wrappers for slightly modified function and the second one changes the whole qemu driver at once. Unfortunately the second patch couldn't be broken up to more parts due to the nature of the fix. Enough chit-chat, let's look at the codes! Martin Kletzander (2): conf: Rework virDomainObjListFindByUUID to allow more concurrent APIs qemu: completely rework reference counting src/conf/domain_conf.c | 27 +- src/conf/domain_conf.h | 2 + src/libvirt_private.syms | 1 + src/qemu/THREADS.txt | 40 ++- src/qemu/qemu_domain.c | 28 +- src/qemu/qemu_domain.h | 12 +- src/qemu/qemu_driver.c | 708 ++++++++++++++++------------------------------ src/qemu/qemu_migration.c | 108 +++---- src/qemu/qemu_migration.h | 10 +- src/qemu/qemu_process.c | 126 ++++----- 10 files changed, 406 insertions(+), 656 deletions(-) -- 2.1.3

3 13

[libvirt] [PATCH v2] qemu: Allow UEFI paths to be specified at compile time
by Michal Privoznik 03 Dec '14

03 Dec '14

Up until now there are just two ways how to specify UEFI paths to libvirt. The first one is editing qemu.conf, the other is editing qemu_conf.c and recompile. Therefore, two new configure options are invented: --with-default-loader --with-default-nvram. At the same time, the list of default paths is extended to AAVMF. which is an aarch64 port of OVMF. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- diff to v1: -introduce configure options too configure.ac | 22 ++++++++++++++++++++++ src/qemu/qemu.conf | 12 +++++++++--- src/qemu/qemu_conf.c | 24 ++++++++++++++++++++---- src/qemu/test_libvirtd_qemu.aug.in | 1 + 4 files changed, 52 insertions(+), 7 deletions(-) diff --git a/configure.ac b/configure.ac index 9fd44b2..c22ac75 100644 --- a/configure.ac +++ b/configure.ac @@ -2786,6 +2786,28 @@ AC_ARG_WITH([default-editor], [DEFAULT_EDITOR=vi]) AC_DEFINE_UNQUOTED([DEFAULT_EDITOR], ["$DEFAULT_EDITOR"], [Default editor to use]) +AC_ARG_WITH([default-loader], + [AS_HELP_STRING([--with-default-loader], + [Default UEFI loader path @<:@default=paths to OVMF and its clones@:>@])], + [with_default_loader=${withval}], + [with_default_loader=no]) +AC_ARG_WITH([default-nvram], + [AS_HELP_STRING([--with-default-nvram], + [Default UEFI NVRAM path @<:@default=paths to OVMF and its clones@:>@])], + [with_default_nvram=${withval}], + [with_default_nvram=no]) + +if test "$with_default_loader:$with_default_nvram" != "no:no"; then + if test "$with_default_loader" = "no"; then + AC_MSG_ERROR([Can't use nvram without loader]) + fi + if test "$with_default_nvram" = "no"; then + AC_MSG_ERROR([Can't use loader without nvram]) + fi + AC_DEFINE_UNQUOTED([DEFAULT_LOADER], ["$with_default_loader"], [Default UEFI loader]) + AC_DEFINE_UNQUOTED([DEFAULT_NVRAM], ["$with_default_nvram"], [Default UEFI nvram]) +fi + # Some GNULIB base64 symbols clash with a kerberos library AC_DEFINE_UNQUOTED([isbase64],[libvirt_gl_isbase64],[Hack to avoid symbol clash]) AC_DEFINE_UNQUOTED([base64_encode],[libvirt_gl_base64_encode],[Hack to avoid symbol clash]) diff --git a/src/qemu/qemu.conf b/src/qemu/qemu.conf index c6db568..1c589a2 100644 --- a/src/qemu/qemu.conf +++ b/src/qemu/qemu.conf @@ -506,6 +506,12 @@ # however, have different variables store. Therefore the nvram is # a list of strings when a single item is in form of: # ${PATH_TO_UEFI_FW}:${PATH_TO_UEFI_VARS}. -# Later, when libvirt creates per domain variable store, this -# list is searched for the master image. -#nvram = [ "/usr/share/OVMF/OVMF_CODE.fd:/usr/share/OVMF/OVMF_VARS.fd" ] +# Later, when libvirt creates per domain variable store, this list is +# searched for the master image. The UEFI firmware can be called +# differently for different guest architectures. For instance, it's OVMF +# for x86_64 and i686, but it's AAVMF for aarch64. The libvirt default +# follows this scheme. +#nvram = [ +# "/usr/share/OVMF/OVMF_CODE.fd:/usr/share/OVMF/OVMF_VARS.fd", +# "/usr/share/AAVMF/AAVMF_CODE.fd:/usr/share/AAVMF/AAVMF_VARS.fd" +#] diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 4764bef..d268fc3 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -107,8 +107,10 @@ void qemuDomainCmdlineDefFree(qemuDomainCmdlineDefPtr def) VIR_FREE(def); } -#define VIR_QEMU_LOADER_FILE_PATH "/usr/share/OVMF/OVMF_CODE.fd" -#define VIR_QEMU_NVRAM_FILE_PATH "/usr/share/OVMF/OVMF_VARS.fd" +#define VIR_QEMU_OVMF_LOADER_PATH "/usr/share/AAVMF/AAVMF_CODE.fd" +#define VIR_QEMU_OVMF_NVRAM_PATH "/usr/share/AAVMF/AAVMF_VARS.fd" +#define VIR_QEMU_AAVMF_LOADER_PATH "/usr/share/OVMF/OVMF_CODE.fd" +#define VIR_QEMU_AAVMF_NVRAM_PATH "/usr/share/OVMF/OVMF_VARS.fd" virQEMUDriverConfigPtr virQEMUDriverConfigNew(bool privileged) { @@ -258,15 +260,29 @@ virQEMUDriverConfigPtr virQEMUDriverConfigNew(bool privileged) cfg->logTimestamp = true; +#if defined(DEFAULT_LOADER) && defined(DEFAULT_NVRAM) if (VIR_ALLOC_N(cfg->loader, 1) < 0 || VIR_ALLOC_N(cfg->nvram, 1) < 0) goto error; cfg->nloader = 1; - if (VIR_STRDUP(cfg->loader[0], VIR_QEMU_LOADER_FILE_PATH) < 0 || - VIR_STRDUP(cfg->nvram[0], VIR_QEMU_NVRAM_FILE_PATH) < 0) + if (VIR_STRDUP(cfg->loader[0], DEFAULT_LOADER) < 0 || + VIR_STRDUP(cfg->nvram[0], DEFAULT_NVRAM) < 0) goto error; +#else /* defined(DEFAULT_LOADER && defined(DEFAULT_NVRAM) */ + + if (VIR_ALLOC_N(cfg->loader, 2) < 0 || + VIR_ALLOC_N(cfg->nvram, 2) < 0) + goto error; + cfg->nloader = 2; + + if (VIR_STRDUP(cfg->loader[0], VIR_QEMU_OVMF_LOADER_PATH) < 0 || + VIR_STRDUP(cfg->nvram[0], VIR_QEMU_OVMF_NVRAM_PATH) < 0 || + VIR_STRDUP(cfg->loader[1], VIR_QEMU_AAVMF_LOADER_PATH) < 0 || + VIR_STRDUP(cfg->nvram[1], VIR_QEMU_AAVMF_NVRAM_PATH) < 0) + goto error; +#endif /* defined(DEFAULT_LOADER && defined(DEFAULT_NVRAM) */ return cfg; error: diff --git a/src/qemu/test_libvirtd_qemu.aug.in b/src/qemu/test_libvirtd_qemu.aug.in index 30fd27e..fc4935b 100644 --- a/src/qemu/test_libvirtd_qemu.aug.in +++ b/src/qemu/test_libvirtd_qemu.aug.in @@ -76,4 +76,5 @@ module Test_libvirtd_qemu = { "log_timestamp" = "0" } { "nvram" { "1" = "/usr/share/OVMF/OVMF_CODE.fd:/usr/share/OVMF/OVMF_VARS.fd" } + { "2" = "/usr/share/AAVMF/AAVMF_CODE.fd:/usr/share/AAVMF/AAVMF_VARS.fd" } } -- 2.0.4

5 12

[libvirt] [PATCH] virsh: vol-upload disallow negative offset
by Shanzhi Yu 03 Dec '14

03 Dec '14

In commit 570d0f, it should diable negative offset both in cmdVolDownload and cmdVolUpload according the description, while it really didn't in cmdVolUpload Disable it in this patch Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1087104 Signed-off-by: Shanzhi Yu <shyu(a)redhat.com> --- tools/virsh-volume.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tools/virsh-volume.c b/tools/virsh-volume.c index 27bd81d..d585ee2 100644 --- a/tools/virsh-volume.c +++ b/tools/virsh-volume.c @@ -677,13 +677,13 @@ cmdVolUpload(vshControl *ctl, const vshCmd *cmd) const char *name = NULL; unsigned long long offset = 0, length = 0; - if (vshCommandOptULongLongWrap(cmd, "offset", &offset) < 0) { - vshError(ctl, _("Unable to parse integer")); + if (vshCommandOptULongLong(cmd, "offset", &offset) < 0) { + vshError(ctl, _("Unable to parse offset value")); return false; } if (vshCommandOptULongLongWrap(cmd, "length", &length) < 0) { - vshError(ctl, _("Unable to parse integer")); + vshError(ctl, _("Unable to parse length value")); return false; } -- 2.1.0

2 1

[libvirt] [PATCH 0/2] fix attached vm cannot get a right label
by Luyao Huang 03 Dec '14

03 Dec '14

When call qemuProcessAttach to attach a qemu process, libvirt will generate a wrong label for DAC, and do not set imagelabel for both of them, no imagelabel will cause some other issue. After this patch guest label will be : <seclabel type='static' model='selinux' relabel='yes'> <label>unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023</label> <imagelabel>system_u:object_r:svirt_image_t:s0-s0:c0.c1023</imagelabel> </seclabel> <seclabel type='static' model='dac' relabel='yes'> <label>+0:+0</label> <imagelabel>+0:+0</imagelabel> </seclabel> Luyao Huang (2): qemu: fix some small issue in qemuProcessAttach security: Add a new func use stat to get process DAC label src/qemu/qemu_process.c | 10 ++++++--- src/security/security_dac.c | 50 +++++++++++++++++++++++++++++++++++++++++++-- 2 files changed, 55 insertions(+), 5 deletions(-) -- 1.8.3.1

2 11

[libvirt] [PATCH] lxc: give RW access to /proc/sys/net/ipv[46] to containers
by Cédric Bosdonnat 03 Dec '14

03 Dec '14

Some programs want to change some values for the network interfaces configuration in /proc/sys/net/ipv[46] folders. Giving RW access on them allows wicked to work on openSUSE 13.2+. In order to mount those folders RW but keep the rest of /proc/sys RO, we add temporary mounts for these folders before bind-mounting /proc/sys. --- src/lxc/lxc_container.c | 97 +++++++++++++++++++++++++++++++++++++++++-------- 1 file changed, 82 insertions(+), 15 deletions(-) diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index 3b08b86..8eb2547 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -800,15 +800,20 @@ typedef struct { int mflags; bool skipUserNS; bool skipUnmounted; + bool temporary; } virLXCBasicMountInfo; static const virLXCBasicMountInfo lxcBasicMounts[] = { - { "proc", "/proc", "proc", MS_NOSUID|MS_NOEXEC|MS_NODEV, false, false }, - { "/proc/sys", "/proc/sys", NULL, MS_BIND|MS_RDONLY, false, false }, - { "sysfs", "/sys", "sysfs", MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, false, false }, - { "securityfs", "/sys/kernel/security", "securityfs", MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, true, true }, + { "proc", "/proc", "proc", MS_NOSUID|MS_NOEXEC|MS_NODEV, false, false, false }, + { "/proc/sys/net/ipv4", "TMP1", NULL, MS_BIND, false, false, true }, + { "/proc/sys/net/ipv6", "TMP2", NULL, MS_BIND, false, false, true }, + { "/proc/sys", "/proc/sys", NULL, MS_BIND|MS_RDONLY, false, false, false }, + { "TMP1", "/proc/sys/net/ipv4", NULL, MS_BIND, false, false, false }, + { "TMP2", "/proc/sys/net/ipv6", NULL, MS_BIND, false, false, false }, + { "sysfs", "/sys", "sysfs", MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, false, false, false }, + { "securityfs", "/sys/kernel/security", "securityfs", MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, true, true, false }, #if WITH_SELINUX - { SELINUX_MOUNT, SELINUX_MOUNT, "selinuxfs", MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, true, true }, + { SELINUX_MOUNT, SELINUX_MOUNT, "selinuxfs", MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, true, true, false }, #endif }; @@ -885,14 +890,23 @@ static int lxcContainerSetReadOnly(void) static int lxcContainerMountBasicFS(bool userns_enabled, bool netns_disabled) { - size_t i; + size_t i, j; int rc = -1; char* mnt_src = NULL; + char* mnt_dst = NULL; int mnt_mflags; + char **tmpkeys = NULL; + char **tmppaths = NULL; + size_t nmounts = ARRAY_CARDINALITY(lxcBasicMounts); VIR_DEBUG("Mounting basic filesystems"); - for (i = 0; i < ARRAY_CARDINALITY(lxcBasicMounts); i++) { + if (VIR_ALLOC_N(tmpkeys, nmounts) < 0 || + VIR_ALLOC_N(tmppaths, nmounts) < 0) { + goto cleanup; + } + + for (i = 0; i < nmounts; i++) { bool bindOverReadonly; virLXCBasicMountInfo const *mnt = &lxcBasicMounts[i]; @@ -906,11 +920,36 @@ static int lxcContainerMountBasicFS(bool userns_enabled, goto cleanup; mnt_mflags = MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY|MS_BIND; } else { - if (VIR_STRDUP(mnt_src, mnt->src) < 0) + /* Look for potential temporary folder match */ + for (j = 0; j < i; j++) { + if (STREQ_NULLABLE(mnt->src, tmpkeys[j])) { + if (VIR_STRDUP(mnt_src, tmppaths[j]) < 0) + goto cleanup; + break; + } + } + if (!mnt_src && VIR_STRDUP(mnt_src, mnt->src) < 0) goto cleanup; mnt_mflags = mnt->mflags; } + if (mnt->temporary) { + char tmppath[] = "/tmp/mount-XXXXXX"; + if (mkdtemp(tmppath) == NULL) { + virReportSystemError(errno, + _("Failed to create temporary folder %s"), + tmppath); + } + if (VIR_STRDUP(tmppaths[i], tmppath) < 0 || + VIR_STRDUP(tmpkeys[i], mnt->dst) < 0 || + VIR_STRDUP(mnt_dst, tmppath) < 0) { + goto cleanup; + } + } else { + if (VIR_STRDUP(mnt_dst, mnt->dst) < 0) + goto cleanup; + } + VIR_DEBUG("Processing %s -> %s", mnt_src, mnt->dst); @@ -940,10 +979,18 @@ static int lxcContainerMountBasicFS(bool userns_enabled, continue; } - if (virFileMakePath(mnt->dst) < 0) { + /* Some source files or folders may not exist like /proc/sys/net/ipv6 + * as they may depend on a kernel module being loaded. */ + if (STRPREFIX(mnt_src, "/") && virFileExists(mnt_src)) { + VIR_DEBUG("Skipping: missing %s", mnt_src); + VIR_FREE(mnt_src); + continue; + } + + if (virFileMakePath(mnt_dst) < 0) { virReportSystemError(errno, _("Failed to mkdir %s"), - mnt_src); + mnt_dst); goto cleanup; } @@ -957,32 +1004,52 @@ static int lxcContainerMountBasicFS(bool userns_enabled, bindOverReadonly = !!(mnt_mflags & MS_RDONLY); VIR_DEBUG("Mount %s on %s type=%s flags=%x", - mnt_src, mnt->dst, mnt->type, mnt_mflags & ~MS_RDONLY); - if (mount(mnt_src, mnt->dst, mnt->type, mnt_mflags & ~MS_RDONLY, NULL) < 0) { + mnt_src, mnt_dst, mnt->type, mnt_mflags & ~MS_RDONLY); + if (mount(mnt_src, mnt_dst, mnt->type, mnt_mflags & ~MS_RDONLY, NULL) < 0) { virReportSystemError(errno, _("Failed to mount %s on %s type %s flags=%x"), - mnt_src, mnt->dst, NULLSTR(mnt->type), + mnt_src, mnt_dst, NULLSTR(mnt->type), mnt_mflags & ~MS_RDONLY); goto cleanup; } if (bindOverReadonly && - mount(mnt_src, mnt->dst, NULL, + mount(mnt_src, mnt_dst, NULL, MS_BIND|MS_REMOUNT|MS_RDONLY, NULL) < 0) { virReportSystemError(errno, _("Failed to re-mount %s on %s flags=%x"), - mnt_src, mnt->dst, + mnt_src, mnt_dst, MS_BIND|MS_REMOUNT|MS_RDONLY); goto cleanup; } VIR_FREE(mnt_src); + VIR_FREE(mnt_dst); } rc = 0; cleanup: + /* Cleanup temporary mounts */ + for (i = 0; i < nmounts; i++) { + virLXCBasicMountInfo const *mnt = &lxcBasicMounts[i]; + if (mnt->temporary) { + if (umount(tmppaths[i]) < 0) { + virReportSystemError(errno, + _("Failed to un-mount temporary %s"), + tmppaths[i]); + } + if (virFileDeleteTree(tmppaths[i]) < 0) + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Failed to remove temporary folder %s"), + tmppaths[i]); + } + } + + virStringFreeList(tmpkeys); + virStringFreeList(tmppaths); VIR_FREE(mnt_src); + VIR_FREE(mnt_dst); VIR_DEBUG("rc=%d", rc); return rc; } -- 2.1.2

2 1

[libvirt] [PATCH] libvirt.spec.in: BuildRequires: python-devel
by Michal Privoznik 03 Dec '14

03 Dec '14

As of f48ab7d8 we are trying to use python-devel to check python's version (or at least gnulib is doing that). However, the specfile doesn't say anything about it and therefore bootstrapping may fail on a minimal installation on rawhide. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- libvirt.spec.in | 1 + 1 file changed, 1 insertion(+) diff --git a/libvirt.spec.in b/libvirt.spec.in index 0959483..083612c 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -424,6 +424,7 @@ BuildRequires: automake BuildRequires: gettext-devel BuildRequires: libtool BuildRequires: /usr/bin/pod2man +BuildRequires: python-devel %endif BuildRequires: git BuildRequires: perl -- 2.0.4

3 4

[libvirt] ANNOUNCE: virt-manager 1.1.0 released
by Cole Robinson 03 Dec '14

03 Dec '14

I'm happy to announce the release of virt-manager 1.1.0! virt-manager is a desktop application for managing KVM, Xen, and LXC virtualization via libvirt. The release can be downloaded from: http://virt-manager.org/download/ The direct download links are: http://virt-manager.org/download/sources/virt-manager/virt-manager-1.1.0.ta… This release includes: - Switch to libosinfo as OS metadata database (Giuseppe Scrivano) - Use libosinfo for OS detection from CDROM media labels (Giuseppe Scrivano) - Use libosinfo for improved OS defaults, like recommended disk size (Giuseppe Scrivano) - virt-image tool has been removed, as previously announced - Enable Hyper-V enlightenments for Windows VMs - Revert virtio-console default, back to plain serial console - Experimental q35 option in new VM 'customize' dialog - UI for virtual network QoS settings (Giuseppe Scrivano) - virt-install: --disk discard= support (Jim Minter) - addhardware: Add spiceport UI (Marc-André Lureau) - virt-install: --events on_poweroff etc. support (Chen Hanxiao) - cli --network portgroup= support and UI support - cli --boot initargs= and UI support - addhardware: allow setting controller model (Chen Hanxiao) - virt-install: support setting hugepage options (Chen Hanxiao) Thanks to everyone who has contributed to this release through testing, bug reporting, submitting patches, and otherwise sending in feedback! Thanks, Cole

2 7

[libvirt] [PATCH v3 0/4] Precreate storage on migration
by Michal Privoznik 03 Dec '14

03 Dec '14

diff to v2: -nits found by Peter fixed Michal Privoznik (4): storage: Introduce storagePoolLookupByTargetPath qemuMonitorJSONBlockStatsUpdateCapacity: Don't skip disks qemu_migration: Send disk sizes to the other side qemu_migration: Precreate missing storage src/libvirt-domain.c | 3 + src/qemu/qemu_migration.c | 319 ++++++++++++++++++++++++++++++++++++++++--- src/qemu/qemu_monitor_json.c | 14 +- src/storage/storage_driver.c | 42 ++++++ src/storage/storage_driver.h | 4 + 5 files changed, 358 insertions(+), 24 deletions(-) -- 2.0.4

2 8

[libvirt] [PATCH 00/11] Replace virXXXFree calls with virObjectUnref
by John Ferlan 03 Dec '14

03 Dec '14

Based on some recent review comments and a bit of internal IRC, this set of patches will change all the existing virXXXFree calls found in daemon/* and src/* to be virObjectUnref instead and then add a rule to inhibit usage unless the string/call is found in docs/*, tests/*, examples/*, tools/*, cfg.mk, libvirt_public.syms, include/libvirt/libvirt-*.h, and src/libvirt-*.c Most of this was brute force to start with and adding the rule afterwards caught a few oddball places. The reason for not wanting to call a virXXXFree function is because it will reset the last error, potentially clearing something and resulting in a message "an error was encountered but the cause is unknown". There were some places in the code which saved the last error, called the free function, then restored the last error. Those have now been adjusted to avoid that processing. Anything that required checking for a non NULL pointer prior to calling the virXXXFree is now not necessary since the virObjectUnref will check for NULL before continuing; whereas, the virXXXFree functions didn't allow NULL. If usage of the virXXXFree function is found during syntax-check, a message such as the following will be displayed: include/libvirt/libvirt-storage.h:256:int virStoragePoolFree (virStoragePoolPtr pool); include/libvirt/libvirt-storage.h:336:int virStorageVolFree (virStorageVolPtr vol); maint.mk: avoid using virXXXFree, use virObjectUnref instead make: *** [sc_prohibit_virXXXFree] Error 1 John Ferlan (11): rpc: Replace virXXXFree with virObjectUnref Replace virDomainFree with virObjectUnref Replace virNetworkFree with virObjectUnref Replace virNodeDeviceFree with virObjectUnref Replace virStorageVolFree with virObjectUnref Replace virStoragePoolFree with virObjectUnref Replace virStreamFree with virObjectUnref Replace virSecretFree with virObjectUnref Replace virNWFilterFree with virObjectUnref Replace virInterfaceFree with virObjectUnref Replace virDomainSnapshotFree with virObjectUnref cfg.mk | 12 +++ daemon/remote.c | 168 ++++++++++++-------------------- daemon/stream.c | 2 +- src/conf/domain_event.c | 4 +- src/conf/network_conf.c | 6 +- src/conf/network_event.c | 2 +- src/conf/node_device_conf.c | 6 +- src/conf/storage_conf.c | 6 +- src/conf/virchrdev.c | 4 +- src/esx/esx_driver.c | 2 +- src/fdstream.c | 2 +- src/hyperv/hyperv_driver.c | 2 +- src/interface/interface_backend_netcf.c | 6 +- src/interface/interface_backend_udev.c | 2 +- src/libxl/libxl_conf.c | 7 +- src/locking/sanlock_helper.c | 3 +- src/lxc/lxc_driver.c | 8 +- src/lxc/lxc_process.c | 8 +- src/nwfilter/nwfilter_driver.c | 6 +- src/qemu/qemu_command.c | 8 +- src/qemu/qemu_driver.c | 4 +- src/qemu/qemu_hotplug.c | 8 +- src/remote/remote_driver.c | 76 +++++++-------- src/rpc/gendispatch.pl | 17 ++-- src/secret/secret_driver.c | 6 +- src/storage/storage_backend.c | 4 +- src/storage/storage_backend_fs.c | 4 +- src/storage/storage_backend_rbd.c | 3 +- src/storage/storage_driver.c | 20 +--- src/uml/uml_conf.c | 2 +- src/vbox/vbox_common.c | 6 +- src/xenconfig/xen_common.c | 2 +- src/xenconfig/xen_sxpr.c | 2 +- 33 files changed, 155 insertions(+), 263 deletions(-) -- 1.9.3

3 24

[libvirt] [PATCH] qemu: update disk chain in vm->newDef when block job compeleted
by weiwei li 03 Dec '14

03 Dec '14

A bug will be hitted with libvirt+qemu+kvm when doing follow steps: 1.initial status :vm = base+z(z is a snapshot) 2. call snapshotCreateXML( ): vm = base+z +z' 3. call blockRebase( ): vm = base+z''(z rebase to z') 4.(after rebasing work done) vm shutdown and then start an error occured like this: "unable to set user and group to '0:0' on ' /path/to/disk': No such file or directory A second start may sucess. The error message reported by virSecurityDACSetSecurityDiskLabel when the disk chain is misused(that is base+z +z', after rebase, but z' was deleted yet).virSecurityDACSetSecurityDiskLabel used an old disk chain after rebasing. In qemuProcessHandleBlockJob when rebase complete, the disk chain in vm->def is updated but vm->newDef is not(when vm->newDef is not null). In shutdown( actrully in qemuProcessStop ), if vm->newDef is not null , vm->def will be replaced by vm->newDef, which contains a wrong disk chain!!! Then the next start will failed with above message. To fix it, update disk chain in vm->newDef in qemuProcessHandleBlockJob. Signed-off-by: Weiwei Li <nuonuoli(a)tencent.com> --- src/qemu/qemu_process.c | 27 +++++++++++++++++++++++++++ 1 files changed, 27 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 24e5f0f..0491249 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -1033,6 +1033,7 @@ qemuProcessHandleBlockJob(qemuMonitorPtr mon ATTRIBUTE_UNUSED, virDomainDiskDefPtr disk; virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver); virDomainDiskDefPtr persistDisk = NULL; + virDomainDiskDefPtr newDisk = NULL; bool save = false; virObjectLock(vm); @@ -1101,6 +1102,32 @@ qemuProcessHandleBlockJob(qemuMonitorPtr mon ATTRIBUTE_UNUSED, disk->mirrorJob = VIR_DOMAIN_BLOCK_JOB_TYPE_UNKNOWN; ignore_value(qemuDomainDetermineDiskChain(driver, vm, disk, true, true)); + /* Chang vm->newDef disk chain */ + if (!persistDisk && vm->newDef){ + int indx = virDomainDiskIndexByName(vm->newDef, + disk->dst, + false); + virStorageSourcePtr copy = NULL; + + if (indx >=0){ + newDisk = vm->newDef->disks[indx]; + copy = virStorageSourceCopy(disk->src, false); + if (virStorageSourceInitChainElement(copy, + newDisk->src, + false) < 0) { + VIR_WARN("Unable to update newDef definition " + "on vm %s after block job", + vm->def->name); + virStorageSourceFree(copy); + copy = NULL; + newDisk = NULL; + } + } + if (copy){ + virStorageSourceFree(newDisk->src); + newDisk->src = copy; + } + } } else if (disk->mirror && (type == VIR_DOMAIN_BLOCK_JOB_TYPE_COPY || type == VIR_DOMAIN_BLOCK_JOB_TYPE_ACTIVE_COMMIT)) { -- 1.7.1

2 2

[libvirt] [PATCH] parallels: fix compilation of parallels_storage.c
by Dmitry Guryanov 02 Dec '14

02 Dec '14

virConnect.privateData is void *, so we can't access fields of parallelsConn, pointer to which is stored in virConnect.privateData. So replace all occurences of conn->privateData->storageState with privconn->storageState. Signed-off-by: Dmitry Guryanov <dguryanov(a)parallels.com> --- src/parallels/parallels_storage.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/src/parallels/parallels_storage.c b/src/parallels/parallels_storage.c index e1b6ea8..882cbd0 100644 --- a/src/parallels/parallels_storage.c +++ b/src/parallels/parallels_storage.c @@ -72,8 +72,8 @@ parallelsStorageClose(virConnectPtr conn) { parallelsConnPtr privconn = conn->privateData; - virStorageDriverStatePtr storageState = conn->privateData->storageState; - conn->privateData->storageState = NULL; + virStorageDriverStatePtr storageState = privconn->storageState; + privconn->storageState = NULL; parallelsStorageLock(storageState); virStoragePoolObjListFree(&privconn->pools); @@ -189,7 +189,7 @@ parallelsPoolCreateByPath(virConnectPtr conn, const char *path) if (!(pool = virStoragePoolObjAssignDef(pools, def))) goto error; - if (virStoragePoolObjSaveDef(conn->privateData->storageState, pool, def) < 0) { + if (virStoragePoolObjSaveDef(privconn->storageState, pool, def) < 0) { virStoragePoolObjRemove(pools, pool); goto error; } @@ -404,7 +404,7 @@ parallelsPoolsAdd(virDomainObjPtr dom, static int parallelsLoadPools(virConnectPtr conn) { parallelsConnPtr privconn = conn->privateData; - virStorageDriverStatePtr storageState = conn->privateData->storageState; + virStorageDriverStatePtr storageState = privconn->storageState; char *base = NULL; size_t i; @@ -461,6 +461,7 @@ parallelsStorageOpen(virConnectPtr conn, virConnectAuthPtr auth ATTRIBUTE_UNUSED, unsigned int flags) { + parallelsConnPtr privconn = conn->privateData; virStorageDriverStatePtr storageState; virCheckFlags(VIR_CONNECT_RO, VIR_DRV_OPEN_ERROR); @@ -475,7 +476,7 @@ parallelsStorageOpen(virConnectPtr conn, return VIR_DRV_OPEN_ERROR; } - conn->privateData->storageState = storageState; + privconn->storageState = storageState; parallelsStorageLock(storageState); if (parallelsLoadPools(conn)) @@ -728,7 +729,7 @@ parallelsStoragePoolDefineXML(virConnectPtr conn, if (!(pool = virStoragePoolObjAssignDef(&privconn->pools, def))) goto cleanup; - if (virStoragePoolObjSaveDef(conn->privateData->storageState, pool, def) < 0) { + if (virStoragePoolObjSaveDef(privconn->storageState, pool, def) < 0) { virStoragePoolObjRemove(&privconn->pools, pool); def = NULL; goto cleanup; -- 1.9.3

2 1

[libvirt] [PATCH 0/2] Simplify MAC address generation on update
by Ján Tomko 02 Dec '14

02 Dec '14

https://bugzilla.redhat.com/show_bug.cgi?id=1156367 Ján Tomko (2): Silently ignore MAC in NetworkLoadConfig Generate a MAC when loading a config instead of package update libvirt.spec.in | 42 ------------------------------------------ src/conf/network_conf.c | 8 ++++++++ 2 files changed, 8 insertions(+), 42 deletions(-) -- 2.0.4

2 4

[libvirt] [PATCHv3 00/16] Network configuration for lxc containers
by Cédric Bosdonnat 02 Dec '14

02 Dec '14

Hi all, Here is a rebased version of v2. Nothing changed except the 'since' version number in the added doc that has been updated. -- Cedric Cédric Bosdonnat (16): Forgot to cleanup ifname_guest* in domain network def parsing Domain conf: allow more than one IP address for net devices LXC: set IP addresses to veth devices in the container lxc conf2xml: convert IP addresses Allow network capabilities hostdev to configure IP addresses lxc conf2xml: convert ip addresses for hostdev NICs Domain network devices can now have a <gateway> element lxc conf2xml: convert lxc.network.ipv[46].gateway LXC: use the new net devices gateway definition LXC: honour network devices link state Wrong place for virDomainNetIpsFormat virNetDevSetIPv4Address: libnl implementation Renamed virNetDevSetIPv4Address to virNetDevSetIPAddress virNetDevAddRoute: implementation using netlink virNetDevClearIPv4Address: netlink implementation Renamed virNetDevClearIPv4Address to virNetDevClearIPAddress docs/formatdomain.html.in | 39 +++ docs/schemas/domaincommon.rng | 55 +++- src/conf/domain_conf.c | 214 +++++++++++++-- src/conf/domain_conf.h | 22 +- src/libvirt_private.syms | 7 +- src/lxc/lxc_container.c | 74 ++++- src/lxc/lxc_native.c | 173 ++++++++---- src/network/bridge_driver.c | 4 +- src/openvz/openvz_conf.c | 2 +- src/openvz/openvz_driver.c | 6 +- src/qemu/qemu_driver.c | 25 +- src/qemu/qemu_hotplug.c | 6 +- src/uml/uml_conf.c | 2 +- src/util/virnetdev.c | 305 ++++++++++++++++++--- src/util/virnetdev.h | 12 +- src/util/virnetlink.c | 38 +++ src/util/virnetlink.h | 2 + src/vbox/vbox_common.c | 3 +- src/xenconfig/xen_common.c | 15 +- src/xenconfig/xen_sxpr.c | 12 +- .../lxcconf2xmldata/lxcconf2xml-physnetwork.config | 4 + tests/lxcconf2xmldata/lxcconf2xml-physnetwork.xml | 3 + tests/lxcconf2xmldata/lxcconf2xml-simple.config | 4 + tests/lxcconf2xmldata/lxcconf2xml-simple.xml | 3 + tests/lxcxml2xmldata/lxc-hostdev.xml | 3 + tests/lxcxml2xmldata/lxc-idmap.xml | 3 + 26 files changed, 880 insertions(+), 156 deletions(-) -- 1.8.4.5

6 39

[libvirt] [PATCH v2 0/4] Precreate storage on migration
by Michal Privoznik 02 Dec '14

02 Dec '14

Version two of: https://www.redhat.com/archives/libvir-list/2014-November/msg01048.html diff to v1: -Switch from qemuDomainGetBlockInfoImpl() to qemuMonitorBlockStatsUpdateCapacity() -s/alloc/capacity/ -small fixes raised by Peter Michal Privoznik (4): storage: Introduce storagePoolLookupByPath qemuMonitorJSONBlockStatsUpdateCapacity: Don't skip disks qemu_migration: Send disk sizes to the other side qemu_migration: Precreate missing storage src/qemu/qemu_migration.c | 313 ++++++++++++++++++++++++++++++++++++++++--- src/qemu/qemu_monitor_json.c | 12 +- src/storage/storage_driver.c | 36 +++++ src/storage/storage_driver.h | 4 + 4 files changed, 342 insertions(+), 23 deletions(-) -- 2.0.4

2 8

[libvirt] [PATCH] nwfilter: fix crash when adding non-existing nwfilter
by Pavel Hrdina 02 Dec '14

02 Dec '14

Adding non-existing nwfilter to a network interface device without any nwfilter specified will crash libvirt daemon with segfault. The reason is that the nwfilter is not found an libvirt will try to restore old nwfilter configuration but there is no nwfilter specified. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/conf/nwfilter_conf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/conf/nwfilter_conf.c b/src/conf/nwfilter_conf.c index 074d745..317792e 100644 --- a/src/conf/nwfilter_conf.c +++ b/src/conf/nwfilter_conf.c @@ -2778,7 +2778,7 @@ virNWFilterObjFindByName(virNWFilterObjListPtr nwfilters, const char *name) for (i = 0; i < nwfilters->count; i++) { virNWFilterObjLock(nwfilters->objs[i]); - if (STREQ(nwfilters->objs[i]->def->name, name)) + if (STREQ_NULLABLE(nwfilters->objs[i]->def->name, name)) return nwfilters->objs[i]; virNWFilterObjUnlock(nwfilters->objs[i]); } -- 2.0.4

2 2

[libvirt] [PATCH] qemu: add tun devices to cgroup ACL
by rongqing.li＠windriver.com 02 Dec '14

02 Dec '14

From: Roy Li <rongqing.li(a)windriver.com> When systemd is used as init, CGroup will be enabled by default, and lead to virsh unable to start VM(which is using the tun/tap), since virsh can not access /dev/net/tun device Signed-off-by: Roy Li <rongqing.li(a)windriver.com> --- src/qemu/qemu.conf | 3 ++- src/qemu/qemu_cgroup.c | 2 +- src/qemu/test_libvirtd_qemu.aug.in | 1 + 3 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu.conf b/src/qemu/qemu.conf index c6db568..225cd0f 100644 --- a/src/qemu/qemu.conf +++ b/src/qemu/qemu.conf @@ -272,7 +272,8 @@ # "/dev/null", "/dev/full", "/dev/zero", # "/dev/random", "/dev/urandom", # "/dev/ptmx", "/dev/kvm", "/dev/kqemu", -# "/dev/rtc","/dev/hpet", "/dev/vfio/vfio" +# "/dev/rtc","/dev/hpet", "/dev/vfio/vfio", +# "/dev/net/tun" #] # # RDMA migration requires the following extra files to be added to the list: diff --git a/src/qemu/qemu_cgroup.c b/src/qemu/qemu_cgroup.c index 0e94cae..7c15c07 100644 --- a/src/qemu/qemu_cgroup.c +++ b/src/qemu/qemu_cgroup.c @@ -45,7 +45,7 @@ static const char *const defaultDeviceACL[] = { "/dev/random", "/dev/urandom", "/dev/ptmx", "/dev/kvm", "/dev/kqemu", "/dev/rtc", "/dev/hpet", "/dev/vfio/vfio", - NULL, + "/dev/net/tun", NULL, }; #define DEVICE_PTY_MAJOR 136 #define DEVICE_SND_MAJOR 116 diff --git a/src/qemu/test_libvirtd_qemu.aug.in b/src/qemu/test_libvirtd_qemu.aug.in index 30fd27e..2ab7a8d 100644 --- a/src/qemu/test_libvirtd_qemu.aug.in +++ b/src/qemu/test_libvirtd_qemu.aug.in @@ -48,6 +48,7 @@ module Test_libvirtd_qemu = { "9" = "/dev/rtc" } { "10" = "/dev/hpet" } { "11" = "/dev/vfio/vfio" } + { "12" = "/dev/net/tun" } } { "save_image_format" = "raw" } { "dump_image_format" = "raw" } -- 2.1.0

2 1

[libvirt] [PATCH] conf: fix a comment typo in virDomainVideoDefaultRAM
by Wang Rui 02 Dec '14

02 Dec '14

Signed-off-by: Wang Rui <moon.wangrui(a)huawei.com> --- src/conf/domain_conf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index fe9b986..2d81c37 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -10221,7 +10221,7 @@ virDomainVideoDefaultRAM(const virDomainDef *def, return 4 * 1024; case VIR_DOMAIN_VIDEO_TYPE_QXL: - /* QEMU use 64M as the minimal video video memory for qxl device */ + /* QEMU use 64M as the minimal video memory for qxl device */ return 64 * 1024; default: -- 1.7.12.4

2 1