January 2021 - Devel - libvirt List Archives

Release of libvirt-6.5.0
by Daniel Veillard 04 Jan '21

04 Jan '21

Half a day late, but I pushed the 6.5.0 release out, it is as usual available as a signed tarball and source rpms from the server: https://libvirt.org/sources/ I also tagged and pushed the 6.5.0 python bindings that one can find at https://libvirt.org/sources/python/ This release includes a number of new features and some improvement, as well as a crash which had made its way in 6.4.0. It will also be my last release of libvirt after close to 15 years, so expect new releases to be signed by Jiri Denemark from now on. * New Features: - Allow firmware blobs configuration QEMU offers a way to tweak how firmware configures itself and/or provide new configuration blobs. New ``<sysinfo/>`` type is introduced that will hold these new blobs. It's possible to either specify new value as a string or provide a filename which contents then serve as the value. - nodedev: Add ability to create mediated devices Mediated devices can now be created with ``virNodeDeviceCreateXML()``. This functionality requires the ``mdevctl`` utility to be installed. The XML schema for node devices was expanded to support attributes for mediated devices. - QEMU: add TPM Proxy device support libvirt can now create guests using a new device type called "TPM Proxy". The TPM Proxy connects to a TPM Resource Manager present in the host, enabling the guest to run in secure virtual machine mode with the help of an Ultravisor. Adding a TPM Proxy to a pSeries guest brings no security benefits unless the guest is running on a PPC64 host that has Ultravisor and TPM Resource Manager support. Only one TPM Proxy is allowed per guest. A guest using a TPM Proxy device can instantiate another TPM device at the same time. This device is supported only for pSeries guests via the new 'spapr-tpm-proxy' model of the TPM 'passthrough' backend. - virhook: Support hooks placed in several files Running all scripts from directory /etc/libvirt/hooks/<driver>.d in alphabetical order. Hook script in old place will be executed as first for backward compatibility. - qemu: Add support for migratable host-passthrough CPU QEMU 2.12 made it possible for guests to use a migration-friendly version of the host-passthrough CPU. This feature is now exposed by libvirt. * Improvements: - network: Support NAT with IPv6 It's now possible to use ``<nat ipv6="yes"/>`` in a libvirt network. - qemu: Auto-fill NUMA information for incomplete topologies If the NUMA topology is not fully described in the guest XML, libvirt will complete it by putting all unspecified CPUs in the first NUMA node. This is only done in the QEMU binary itself supports disjointed CPU ranges for NUMA nodes. - qemu: Assign hostdev-backed interfaces to PCIe slots All SR-IOV capable devices are PCIe, so when their VFs are assigned to guests they should end up in PCIe slots rather than conventional PCI ones. * Bug fixes: - qemu: fixed crash in ``qemuDomainBlockCommit`` This release fixes a regression which was introduced in libvirt v6.4.0 where libvirtd always crashes when a block commit of a disk is requested. - qemu: fixed zPCI address auto generation on s390 Removes the correlation between the zPCI address attributes uid and fid. Fixes the validation and autogeneration of zPCI address attributes. - qemu: Skip pre-creation of NVMe disks during migration libvirt has no way to create NVMe devices on the target host, so it now just makes sure they exist and let the migration proceed in that case. Thanks everybody for the help on putting this release out, and the gazillion ones before :-) Enjoy and stay safe ! Daniel -- Daniel Veillard | Red Hat Developers Tools http://developer.redhat.com/ veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

5 6

[PATCH] kbase: debuglogs: Fix typo in unprivileged libvirtd config path
by Tomáš Janoušek 04 Jan '21

04 Jan '21

Signed-off-by: Tomáš Janoušek <tomi(a)nomi.cz> --- docs/kbase/debuglogs.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/kbase/debuglogs.rst b/docs/kbase/debuglogs.rst index f3937bc337..3fb06bedc3 100644 --- a/docs/kbase/debuglogs.rst +++ b/docs/kbase/debuglogs.rst @@ -68,7 +68,7 @@ variable: However, when you are using the session mode ``qemu:///session`` or you run the ``libvirtd`` as unprivileged user you will find configuration file under -``$XDG_CONFIG_HOME/libvirt/libvirt.conf``. +``$XDG_CONFIG_HOME/libvirt/libvirtd.conf``. Runtime setting ^^^^^^^^^^^^^^^ -- Tomáš Janoušek, a.k.a. Pivník, a.k.a. Liskni_si, https://work.lisk.in/

2 1

[PATCH] qemu: Drop has_ccw_address from _qemuAgentDiskAddress
by Michal Privoznik 04 Jan '21

04 Jan '21

In recent patches new mambers to _qemuAgentDiskAddress struct were introduced to keep optional CCW address sent by the guest agent. These two members are a struct to store CCW address into and a boolean to keep track whether the CCW address is valid. Well, we can hold the same information with a pointer - instead of storing the CCW address structure let's keep just a pointer to it. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/qemu/qemu_agent.c | 18 ++++++++++++------ src/qemu/qemu_agent.h | 3 +-- src/qemu/qemu_driver.c | 9 +++------ 3 files changed, 16 insertions(+), 14 deletions(-) diff --git a/src/qemu/qemu_agent.c b/src/qemu/qemu_agent.c index b155896e78..af0397e6e2 100644 --- a/src/qemu/qemu_agent.c +++ b/src/qemu/qemu_agent.c @@ -1824,6 +1824,7 @@ qemuAgentDiskAddressFree(qemuAgentDiskAddressPtr info) g_free(info->serial); g_free(info->bus_type); g_free(info->devnode); + g_free(info->ccw_addr); g_free(info); } @@ -1899,12 +1900,17 @@ qemuAgentGetDiskAddress(virJSONValuePtr json) GET_DISK_ADDR(pci, &addr->pci_controller.function, "function"); if ((ccw = virJSONValueObjectGet(json, "ccw-address"))) { - addr->has_ccw_address = true; - GET_DISK_ADDR(ccw, &addr->ccw_addr.cssid, "cssid"); - if (addr->ccw_addr.cssid == 0) /* Guest CSSID 0 is 0xfe on host */ - addr->ccw_addr.cssid = 0xfe; - GET_DISK_ADDR(ccw, &addr->ccw_addr.ssid, "ssid"); - GET_DISK_ADDR(ccw, &addr->ccw_addr.devno, "devno"); + g_autofree virDomainDeviceCCWAddressPtr ccw_addr = NULL; + + ccw_addr = g_new0(virDomainDeviceCCWAddress, 1); + + GET_DISK_ADDR(ccw, &ccw_addr->cssid, "cssid"); + if (ccw_addr->cssid == 0) /* Guest CSSID 0 is 0xfe on host */ + ccw_addr->cssid = 0xfe; + GET_DISK_ADDR(ccw, &ccw_addr->ssid, "ssid"); + GET_DISK_ADDR(ccw, &ccw_addr->devno, "devno"); + + addr->ccw_addr = g_steal_pointer(&ccw_addr); } #undef GET_DISK_ADDR diff --git a/src/qemu/qemu_agent.h b/src/qemu/qemu_agent.h index 4ea9b9dc1e..0d47230161 100644 --- a/src/qemu/qemu_agent.h +++ b/src/qemu/qemu_agent.h @@ -77,8 +77,7 @@ struct _qemuAgentDiskAddress { unsigned int target; unsigned int unit; char *devnode; - bool has_ccw_address; - virDomainDeviceCCWAddress ccw_addr; + virDomainDeviceCCWAddressPtr ccw_addr; }; void qemuAgentDiskAddressFree(qemuAgentDiskAddressPtr addr); G_DEFINE_AUTOPTR_CLEANUP_FUNC(qemuAgentDiskAddress, qemuAgentDiskAddressFree); diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 62b0852c33..a376824854 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -18887,8 +18887,7 @@ qemuAgentFSInfoToPublic(qemuAgentFSInfoPtr agent, diskDef = virDomainDiskByAddress(vmdef, &agentdisk->pci_controller, - agentdisk->has_ccw_address ? - &agentdisk->ccw_addr : NULL, + agentdisk->ccw_addr, agentdisk->bus, agentdisk->target, agentdisk->unit); @@ -19931,8 +19930,7 @@ qemuAgentDiskInfoFormatParams(qemuAgentDiskInfoPtr *info, /* match the disk to the target in the vm definition */ diskdef = virDomainDiskByAddress(vmdef, &info[i]->address->pci_controller, - info[i]->address->has_ccw_address ? - &info[i]->address->ccw_addr : NULL, + info[i]->address->ccw_addr, info[i]->address->bus, info[i]->address->target, info[i]->address->unit); @@ -20017,8 +20015,7 @@ qemuAgentFSInfoFormatParams(qemuAgentFSInfoPtr *fsinfo, /* match the disk to the target in the vm definition */ diskdef = virDomainDiskByAddress(vmdef, &d->pci_controller, - d->has_ccw_address ? - &d->ccw_addr : NULL, + d->ccw_addr, d->bus, d->target, d->unit); -- 2.26.2

2 1

live migration is not using secondary interface
by Olaf Hering 04 Jan '21

04 Jan '21

A naive 'virsh migrate --live domU xen+tcp://cross-over-ip' uses the ordinary uplink instead of the requested IP address. According to the documentation an additional option has to be specified to really use the other network interface. However, neither 'tcp://cross-over-ip' nor 'xenmigr:cross-over-ip/' works, the migration does not start. 'xenmigr' was removed with 1dac5fbbbb06a0341e8087dc33af75c8352d77a4 and 7ed598438687feaddaf0a653d7cbb8a1c1ad4933, but the docs were not updated. This leaves 'tcp://ip'. What change needs to be done in the driver to support a secondary interface? Olaf

3 4

[RFC PATCH 0/3] Fix "virsh domfsinfo" on s390x (again)
by Thomas Huth 04 Jan '21

04 Jan '21

My previous attempts to fix "virsh domfsinfo" on s390x were unfortunately wrong due to some misunderstandings on my side. To correctly list the "Target" device in the output of "virsh domfsinfo", we need to search through the available devices using their CCW address on s390x. For this the QEMU guest agent will be enhanced to also send the CCW address of devices in the guest (see the following URL for details: https://lore.kernel.org/qemu-devel/20201125105417.380317-1-thuth@redhat.com/ ... but it has just been posted and will take some time to get merged, since QEMU is still in hard freeze, that's why I've labed this libvirt series here as RFC only). Using this CCW address, we then can look up the correct target devices on the libvirt side, too. See also https://bugzilla.redhat.com/show_bug.cgi?id=1858771 for some more information. Thomas Huth (3): qemu: agent: Store CCW address in qemuAgentDiskInfo if provided by the guest domain_conf: Allow to look up virtio-block devices by their CCW address domain_conf: Allow to look up scsi disks when controller uses a CCW address src/conf/domain_conf.c | 33 ++++++++++++++++++++++++++++++++- src/conf/domain_conf.h | 4 ++++ src/qemu/qemu_agent.c | 11 +++++++++++ src/qemu/qemu_agent.h | 2 ++ src/qemu/qemu_driver.c | 8 +++++--- 5 files changed, 54 insertions(+), 4 deletions(-) -- 2.18.4

3 12

[PATCH] virt-aa-helper: disallow graphics socket read permissions
by Simon Arlott 02 Jan '21

02 Jan '21

The VM does not need read permission for its own sockets to create(), bind(), accept() connections or to recv(), send(), etc. on connections. This was fixed in ab9569e5460d1e4737fe8b625c67687dc2204665 (virt-aa-helper: disallow VNC socket read permissions), but then b6465e1aa49397367a9cd0f27110b9c2280a7385 (graphics: introduce new listen type 'socket') and acc83afe333bfadd3f7f79091d38ca3d7da1eeb2 (acc83afe333bfadd3f7f79091d38ca3d7da1eeb2) reverted it. Unless the read permission is omitted, VMs can connect to each other's VNC/graphics sockets. Signed-off-by: Simon Arlott <libvirt(a)octiron.net> --- src/security/virt-aa-helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/security/virt-aa-helper.c b/src/security/virt-aa-helper.c index 6e6dd1b1db..fddbdafc41 100644 --- a/src/security/virt-aa-helper.c +++ b/src/security/virt-aa-helper.c @@ -1053,7 +1053,7 @@ get_files(vahControl * ctl) if (listenObj.type == VIR_DOMAIN_GRAPHICS_LISTEN_TYPE_SOCKET && listenObj.socket && - vah_add_file(&buf, listenObj.socket, "rw")) + vah_add_file(&buf, listenObj.socket, "w")) goto cleanup; } } -- 2.17.1 -- Simon Arlott

3 6