January 2018 - Devel - libvirt List Archives

[libvirt] [PATCH v4 0/2] qemu: Add support for hot unplugging redirdev
by Chen Hanxiao 05 Jan '18

05 Jan '18

v4: update commit messages and news v3: use helper qemuDomainDelChardevTLSObjects and address some comments v2: rebased on master Chen Hanxiao (2): qemu: Add support for hot unplugging redirdev device which can use the detach-device --live news: add change of hot unplug redirdev docs/news.xml | 5 +++ src/qemu/qemu_driver.c | 4 ++- src/qemu/qemu_hotplug.c | 90 +++++++++++++++++++++++++++++++++++++++++++++++++ src/qemu/qemu_hotplug.h | 4 +++ 4 files changed, 102 insertions(+), 1 deletion(-) -- 2.14.3

1 2

[libvirt] [PATCH v3 0/6] Add setting CPU features (CPUID) with libxenlight driver.
by Marek Marczykowski-Górecki 05 Jan '18

05 Jan '18

Add support for CPUID setting based on <cpu> element. Since libxl format support only adjusting specific bits over host CPU, only mode='host-passthrough' is supported - other values are rejected (including default 'custom'). This will break some configurations working before (bare <cpu> element with for example NUMA configuration), but libxl driver never supported full 'custom' mode - it was silently ignored, which might lead to some unexpected effects. Since mode='host-passthrough' is now necessary to specify CPU options, do not enable nested HVM feature by mere presence of this element, require explicit enabling "vmx" or "svm" bits. Nested HVM is still in "preview" state, so better be explicit here. v2 of this patch series: https://www.redhat.com/archives/libvir-list/2017-July/msg00050.html Marek Marczykowski-Górecki (6): libxl: error out on not supported CPU mode, instead of silently ignoring libxl: do not enable nested HVM by mere presence of <cpu> element libxl: add support for CPUID features policy tests: check CPU features handling in libxl driver xenconfig: add CPUID handling to domXML <-> xl.cfg conversion tests: add test case for CPUID in xenconfig driver src/libxl/libxl_conf.c | 53 +- src/xenconfig/xen_xl.c | 260 ++++++++- src/xenconfig/xen_xl.h | 2 +- tests/libxlxml2domconfigdata/fullvirt-cpuid.json | 64 ++- tests/libxlxml2domconfigdata/fullvirt-cpuid.xml | 37 +- tests/libxlxml2domconfigdata/vnuma-hvm.json | 1 +- tests/libxlxml2domconfigtest.c | 1 +- tests/xlconfigdata/test-fullvirt-cpuid.cfg | 25 +- tests/xlconfigdata/test-fullvirt-cpuid.xml | 36 +- tests/xlconfigdata/test-fullvirt-nestedhvm.xml | 4 +- tests/xlconfigdata/test-fullvirt-vnuma-autocomplete.xml | 2 +- tests/xlconfigdata/test-fullvirt-vnuma-nodistances.xml | 2 +- tests/xlconfigdata/test-fullvirt-vnuma-partialdist.xml | 2 +- tests/xlconfigdata/test-fullvirt-vnuma.xml | 2 +- tests/xlconfigtest.c | 1 +- 15 files changed, 453 insertions(+), 39 deletions(-) create mode 100644 tests/libxlxml2domconfigdata/fullvirt-cpuid.json create mode 100644 tests/libxlxml2domconfigdata/fullvirt-cpuid.xml create mode 100644 tests/xlconfigdata/test-fullvirt-cpuid.cfg create mode 100644 tests/xlconfigdata/test-fullvirt-cpuid.xml base-commit: 984c534a3f2219444f4cb4df61d77b8c6e5054d7 -- git-series 0.9.1

4 22

[libvirt] [PATCH V3] blockjob: Fix error checking of blockjob status
by Jie Wang 05 Jan '18

05 Jan '18

offset and len can also be equal to 0 on failed if blockjob return status:"BLOCK_JOB_COMPLETED" with error:"File descriptor in bad state", so it's better to make our decision based on whether 'error' is non-NULL. --- src/qemu/qemu_monitor_json.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index e45868b..18b2ad6 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -885,8 +885,7 @@ qemuMonitorJSONHandleBlockJobImpl(qemuMonitorPtr mon, switch ((virConnectDomainEventBlockJobStatus) event) { case VIR_DOMAIN_BLOCK_JOB_COMPLETED: error = virJSONValueObjectGetString(data, "error"); - /* Make sure the whole device has been processed */ - if (offset != len) + if (offset != len || error) event = VIR_DOMAIN_BLOCK_JOB_FAILED; break; case VIR_DOMAIN_BLOCK_JOB_CANCELED: -- 1.8.3.1

1 0

[libvirt] [PATCH] qemu_hotplug: use VIR_ERR_NO_DEVICE when target detaching device is not found
by Chen Hanxiao 05 Jan '18

05 Jan '18

From: Chen Hanxiao <chenhanxiao(a)gmail.com> We used VIR_ERR_OPERATION_FAILED when target detaching device is not found. That error code VIR_ERR_OPERATION_FAILED is widely used, so the tools powered by libvirt, such as nova, can't catch the exact errors from libvirt. This patch uses VIR_ERR_NO_DEVICE instead. Signed-off-by: Chen Hanxiao <chenhanxiao(a)gmail.com> --- src/libvirt_private.syms | 2 ++ src/qemu/qemu_hotplug.c | 51 +++++++++++++++++++++++++++++++----------------- 2 files changed, 35 insertions(+), 18 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index d5c3b9abb..31e83f152 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -203,6 +203,7 @@ virDomainChrConsoleTargetTypeToString; virDomainChrDefForeach; virDomainChrDefFree; virDomainChrDefNew; +virDomainChrDeviceTypeToString; virDomainChrEquals; virDomainChrFind; virDomainChrGetDomainPtrs; @@ -427,6 +428,7 @@ virDomainMemoryDefFree; virDomainMemoryFindByDef; virDomainMemoryFindInactiveByDef; virDomainMemoryInsert; +virDomainMemoryModelTypeToString; virDomainMemoryRemove; virDomainMemorySourceTypeFromString; virDomainMemorySourceTypeToString; diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 7de04c85a..0fa3c54c0 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -3454,7 +3454,7 @@ qemuDomainChangeGraphics(virQEMUDriverPtr driver, int ret = -1; if (!olddev) { - virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + virReportError(VIR_ERR_NO_DEVICE, "%s", _("cannot find existing graphics device to modify")); goto cleanup; } @@ -4743,7 +4743,7 @@ int qemuDomainDetachControllerDevice(virQEMUDriverPtr driver, if ((idx = virDomainControllerFind(vm->def, dev->data.controller->type, dev->data.controller->idx)) < 0) { - virReportError(VIR_ERR_OPERATION_FAILED, + virReportError(VIR_ERR_NO_DEVICE, _("controller %s:%d not found"), virDomainControllerTypeToString(dev->data.controller->type), dev->data.controller->idx); @@ -4972,18 +4972,18 @@ int qemuDomainDetachHostDevice(virQEMUDriverPtr driver, if (idx < 0) { switch (subsys->type) { case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI: - virReportError(VIR_ERR_OPERATION_FAILED, + virReportError(VIR_ERR_NO_DEVICE, _("host pci device %.4x:%.2x:%.2x.%.1x not found"), pcisrc->addr.domain, pcisrc->addr.bus, pcisrc->addr.slot, pcisrc->addr.function); break; case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB: if (usbsrc->bus && usbsrc->device) { - virReportError(VIR_ERR_OPERATION_FAILED, + virReportError(VIR_ERR_NO_DEVICE, _("host usb device %03d.%03d not found"), usbsrc->bus, usbsrc->device); } else { - virReportError(VIR_ERR_OPERATION_FAILED, + virReportError(VIR_ERR_NO_DEVICE, _("host usb device vendor=0x%.4x product=0x%.4x not found"), usbsrc->vendor, usbsrc->product); } @@ -4992,13 +4992,13 @@ int qemuDomainDetachHostDevice(virQEMUDriverPtr driver, if (scsisrc->protocol == VIR_DOMAIN_HOSTDEV_SCSI_PROTOCOL_TYPE_ISCSI) { virDomainHostdevSubsysSCSIiSCSIPtr iscsisrc = &scsisrc->u.iscsi; - virReportError(VIR_ERR_OPERATION_FAILED, + virReportError(VIR_ERR_NO_DEVICE, _("host scsi iSCSI path %s not found"), iscsisrc->src->path); } else { virDomainHostdevSubsysSCSIHostPtr scsihostsrc = &scsisrc->u.host; - virReportError(VIR_ERR_OPERATION_FAILED, + virReportError(VIR_ERR_NO_DEVICE, _("host scsi device %s:%u:%u.%llu not found"), scsihostsrc->adapter, scsihostsrc->bus, scsihostsrc->target, scsihostsrc->unit); @@ -5036,8 +5036,10 @@ qemuDomainDetachShmemDevice(virQEMUDriverPtr driver, qemuDomainObjPrivatePtr priv = vm->privateData; if ((idx = virDomainShmemDefFind(vm->def, dev)) < 0) { - virReportError(VIR_ERR_OPERATION_INVALID, "%s", - _("device not present in domain configuration")); + virReportError(VIR_ERR_NO_DEVICE, + _("Shmem device of model '%s' not found " + "in domain configuration"), + virDomainShmemModelTypeToString(dev->model)); return -1; } @@ -5093,8 +5095,10 @@ qemuDomainDetachWatchdog(virQEMUDriverPtr driver, watchdog->model == dev->model && watchdog->action == dev->action && virDomainDeviceInfoAddressIsEqual(&dev->info, &watchdog->info))) { - virReportError(VIR_ERR_OPERATION_INVALID, "%s", - _("watchdog device not present in domain configuration")); + virReportError(VIR_ERR_NO_DEVICE, + _("watchdog device of model '%s' is not " + "found in domain configuration"), + virDomainWatchdogModelTypeToString(watchdog->model)); return -1; } @@ -5134,8 +5138,13 @@ qemuDomainDetachNetDevice(virQEMUDriverPtr driver, virDomainNetDefPtr detach = NULL; qemuDomainObjPrivatePtr priv = vm->privateData; - if ((detachidx = virDomainNetFindIdx(vm->def, dev->data.net)) < 0) + if ((detachidx = virDomainNetFindIdx(vm->def, dev->data.net)) < 0) { + char mac[VIR_MAC_STRING_BUFLEN]; + virReportError(VIR_ERR_NO_DEVICE, + _("netdev '%s' not found in domain configuration"), + virMacAddrFormat(&dev->data.net->mac, mac)); goto cleanup; + } detach = vm->def->nets[detachidx]; @@ -5321,8 +5330,10 @@ int qemuDomainDetachChrDevice(virQEMUDriverPtr driver, char *devstr = NULL; if (!(tmpChr = virDomainChrFind(vmdef, chr))) { - virReportError(VIR_ERR_OPERATION_INVALID, "%s", - _("device not present in domain configuration")); + virReportError(VIR_ERR_NO_DEVICE, + _("Chr device of type '%s' not found " + "in domain configuration"), + virDomainChrDeviceTypeToString(chr->deviceType)); goto cleanup; } @@ -5368,8 +5379,10 @@ qemuDomainDetachRNGDevice(virQEMUDriverPtr driver, int ret = -1; if ((idx = virDomainRNGFind(vm->def, rng)) < 0) { - virReportError(VIR_ERR_OPERATION_INVALID, "%s", - _("device not present in domain configuration")); + virReportError(VIR_ERR_NO_DEVICE, + _("RNG device of model '%s' not found " + "in domain configuration"), + virDomainRNGBackendTypeToString(rng->model)); return -1; } @@ -5411,8 +5424,10 @@ qemuDomainDetachMemoryDevice(virQEMUDriverPtr driver, qemuDomainMemoryDeviceAlignSize(vm->def, memdef); if ((idx = virDomainMemoryFindByDef(vm->def, memdef)) < 0) { - virReportError(VIR_ERR_OPERATION_INVALID, "%s", - _("device not present in domain configuration")); + virReportError(VIR_ERR_NO_DEVICE, + _("memory device of model '%s' not found " + "in domain configuration"), + virDomainMemoryModelTypeToString(memdef->model)); return -1; } -- 2.14.3

2 1

[libvirt] [PATCH V2] blockjob: Fix error checking of blockjob status
by Jie Wang 04 Jan '18

04 Jan '18

offset and len can also be equal to 0 on failed if blockjob return status:"BLOCK_JOB_COMPLETED" with error:"File descriptor in bad state", so it's better to make our decision based on whether 'error' is non-NULL. --- src/qemu/qemu_monitor_json.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index e45868b..ae64a5a 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -885,8 +885,7 @@ qemuMonitorJSONHandleBlockJobImpl(qemuMonitorPtr mon, switch ((virConnectDomainEventBlockJobStatus) event) { case VIR_DOMAIN_BLOCK_JOB_COMPLETED: error = virJSONValueObjectGetString(data, "error"); - /* Make sure the whole device has been processed */ - if (offset != len) + if (error != NULL) event = VIR_DOMAIN_BLOCK_JOB_FAILED; break; case VIR_DOMAIN_BLOCK_JOB_CANCELED: -- 1.8.3.1

2 1

[libvirt] [PATCH v2 0/3] Storage pool common object fixes
by John Ferlan 04 Jan '18

04 Jan '18

v1: https://www.redhat.com/archives/libvir-list/2017-December/msg00543.html Changes since v1... * Added a patch to handle a NULL return with pool obj lock * Alter the IsDuplicate API to use a bool parameter * Use the IsDuplicate API from the test driver. This would have generated the correct error message about a duplicate UUID instead of the Duplicate key that was generated. Ran virt-manager tests prior to Cole's fixes and of course after. John Ferlan (3): conf: Need to unlock pools on object allocation failure conf: Use bool for @check_active parameter test: Use virStoragePoolObjIsDuplicate for storage define/create src/conf/virstorageobj.c | 4 ++-- src/conf/virstorageobj.h | 2 +- src/storage/storage_driver.c | 4 ++-- src/test/test_driver.c | 11 ++++------- 4 files changed, 9 insertions(+), 12 deletions(-) -- 2.13.6

2 6

[libvirt] [PATCH] rpc: remove redundant logic
by Pavel Hrdina 04 Jan '18

04 Jan '18

Introduced by commit <0eaa59dce1>. That comparison already returns true or false. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- Pushed under trivial rule. src/rpc/virnetserverclient.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/rpc/virnetserverclient.c b/src/rpc/virnetserverclient.c index cae5375359..1f335d3b1e 100644 --- a/src/rpc/virnetserverclient.c +++ b/src/rpc/virnetserverclient.c @@ -1090,7 +1090,7 @@ virNetServerClientClose(virNetServerClientPtr client) bool virNetServerClientIsClosedLocked(virNetServerClientPtr client) { - return client->sock == NULL ? true : false; + return client->sock == NULL; } -- 2.14.3

1 0

[libvirt] [PATCH 0/4] Resolve issue with duplicated SCSI hostdev addresses
by John Ferlan 04 Jan '18

04 Jan '18

Details in the various patches. John Ferlan (4): qemu: Tolerate storage source private data being NULL for hotplug SCSI hostdev qemu: Use same model when adding hostdev SCSI controller conf: Use existing SCSI hostdev model to create new conf: Fix generating addresses for SCSI hostdev src/conf/domain_conf.c | 44 ++++++++++++---------- src/qemu/qemu_hotplug.c | 19 +++++++--- .../hostdev-scsi-autogen-address.xml | 2 +- 3 files changed, 39 insertions(+), 26 deletions(-) -- 2.13.6

3 22

[libvirt] Qemu Commandline options are ignored during migration
by Tim Foerster 04 Jan '18

04 Jan '18

Heyho Guys, The domain defintion xml allows to extend the qemu command. For example: ``` <qemu:commandline> <qemu:arg value='-cpu'/> <qemu:arg value='SandyBridge,-kvm_steal_time'/> </qemu:commandline> ``` The options are correctly passed on startup, but the migration will ignore them. The desination xml is dumped with the flags 1|2|8. (secure, inactive and migrateable) `secure` flag is required to keep vnc passwords, and inactive to keep the metadata section. But i also mentioned that qemu:commandline section needs inactive flag too. -- Tim Foerster

1 0

[libvirt] [PATCH] vhost-user: no need to remove chardev when remove netdev
by linzhecheng 04 Jan '18

04 Jan '18

We use virsh detach-device to hot-unplug a vhost-user net card, and libvirtd will issue qmp_netdev_del, qmp_chardev_remove and qmp_netdev_del. However, in qmp_netdev_del, qemu will remove the chardev when cleanup the netdev. So libvirtd has no need to issue qmp_chardev_remove to qemu any more. Change-Id: I4883c17bbb8dce088406f6a8a9ca4e02612d21f4 diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 7de04c85a..c4787dd60 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -4014,7 +4014,6 @@ qemuDomainRemoveNetDevice(virQEMUDriverPtr driver, qemuDomainObjPrivatePtr priv = vm->privateData; virObjectEventPtr event; char *hostnet_name = NULL; - char *charDevAlias = NULL; size_t i; int ret = -1; int actualType = virDomainNetGetActualType(net); @@ -4029,8 +4028,7 @@ qemuDomainRemoveNetDevice(virQEMUDriverPtr driver, VIR_DEBUG("Removing network interface %s from domain %p %s", net->info.alias, vm, vm->def->name); - if (virAsprintf(&hostnet_name, "host%s", net->info.alias) < 0 || - !(charDevAlias = qemuAliasChardevFromDevAlias(net->info.alias))) + if (virAsprintf(&hostnet_name, "host%s", net->info.alias) < 0) goto cleanup; @@ -4057,15 +4055,6 @@ qemuDomainRemoveNetDevice(virQEMUDriverPtr driver, } } - if (actualType == VIR_DOMAIN_NET_TYPE_VHOSTUSER) { - /* vhostuser has a chardev too */ - if (qemuMonitorDetachCharDev(priv->mon, charDevAlias) < 0) { - /* well, this is a messy situation. Guest visible PCI device has - * been removed, netdev too but chardev not. The best seems to be - * to just ignore the error and carry on. - */ - } - } if (qemuDomainObjExitMonitor(driver, vm) < 0) goto cleanup; @@ -4108,7 +4097,6 @@ qemuDomainRemoveNetDevice(virQEMUDriverPtr driver, cleanup: virObjectUnref(cfg); - VIR_FREE(charDevAlias); VIR_FREE(hostnet_name); return ret; } -- 2.12.2.windows.2

2 1