- Devel - libvirt List Archives

[libvirt] [PATCH RFC] lxc: Include support to lxc version 3.0 and higher.
by Julio Faracco 01 Oct '18

01 Oct '18

This patch introduce the new settings for LXC 3.0 and higher. The older versions keep the compatibility to deprecated settings for LXC, but after release 3.0, the compatibility was removed. This commit adds the support to the refactored settings. Signed-off-by: Julio Faracco <jcfaracco(a)gmail.com> --- src/lxc/lxc_native.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/src/lxc/lxc_native.c b/src/lxc/lxc_native.c index cbdec723dd..c0f70ebb7d 100644 --- a/src/lxc/lxc_native.c +++ b/src/lxc/lxc_native.c @@ -200,7 +200,9 @@ lxcSetRootfs(virDomainDefPtr def, int type = VIR_DOMAIN_FS_TYPE_MOUNT; VIR_AUTOFREE(char *) value = NULL; - if (virConfGetValueString(properties, "lxc.rootfs", &value) <= 0) + if (virConfGetValueString(properties, "lxc.rootfs", &value) <= 0 && + /* Legacy config keys were removed after release 3.0. */ + virConfGetValueString(properties, "lxc.rootfs.path", &value) <= 0) return -1; if (STRPREFIX(value, "/dev/")) @@ -1041,7 +1043,9 @@ lxcParseConfigString(const char *config, if (VIR_STRDUP(vmdef->os.init, "/sbin/init") < 0) goto error; - if (virConfGetValueString(properties, "lxc.utsname", &value) <= 0 || + if ((virConfGetValueString(properties, "lxc.utsname", &value) <= 0 && + /* Legacy config keys were removed after release 3.0. */ + virConfGetValueString(properties, "lxc.uts.name", &value) <= 0) || VIR_STRDUP(vmdef->name, value) < 0) goto error; if (!vmdef->name && (VIR_STRDUP(vmdef->name, "unnamed") < 0)) @@ -1051,7 +1055,9 @@ lxcParseConfigString(const char *config, goto error; /* Look for fstab: we shouldn't have it */ - if (virConfGetValue(properties, "lxc.mount")) { + if (virConfGetValue(properties, "lxc.mount") || + /* Legacy config keys were removed after release 3.0. */ + virConfGetValue(properties, "lxc.mount.fstab")) { virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s", _("lxc.mount found, use lxc.mount.entry lines instead")); goto error; -- 2.17.1

2 3

[libvirt] [PATCHv4 00/18] Introduce x86 Cache Monitoring Technology (CMT)
by Wang Huaqiang 01 Oct '18

01 Oct '18

This series of patches and the series already been merged introduce the x86 Cache Monitoring Technology (CMT) to libvirt by interacting with kernel resource control (resctrl) interface. CMT is one of the Intel(R) x86 CPU feature which belongs to the Resource Director Technology (RDT). CMT reports the occupancy of the last level cache, which is shared by all CPU cores. In the v1 series, an original and complete feature for CMT was introduced The v2 and v3 patches address the feature for the host capability of CMT. v4 is addressing the feature for monitoring VM vcpu thread set cache occupancy and reporting it through a virsh command. We have serval discussion about the enabling of CMT, please refer to following links for the RFCs. RFCv3 https://www.redhat.com/archives/libvir-list/2018-August/msg01213.html RFCv2 https://www.redhat.com/archives/libvir-list/2018-July/msg00409.html https://www.redhat.com/archives/libvir-list/2018-July/msg01241.html RFCv1 https://www.redhat.com/archives/libvir-list/2018-June/msg00674.html And the merged commits are list as below, for host capability of CMT. 6af8417415508c31f8ce71234b573b4999f35980 8f6887998bf63594ae26e3db18d4d5896c5f2cb4 58fcee6f3a2b7e89c21c1fb4ec21429c31a0c5b8 12093f1feaf8f5023dcd9d65dff111022842183d a5d293c18831dcf69ec6195798387fbb70c9f461 1. About reason why CMT is necessary in libvirt? The perf events of 'CMT, MBML, MBMT' have been phased out since Linux kernel commit c39a0e2c8850f08249383f2425dbd8dbe4baad69, in libvirt the perf based cmt,mbm will not work with the latest linux kernel. These patches add CMT feature to libvirt through kernel resctrlfs interface. 2 Create cache monitoring group (cache monitor). The main interface for creating monitoring group is through XML file. The proposed configuration is like: <cputune> <cachetune vcpus='1'> <cache id='0' level='3' type='code' size='7680' unit='KiB'/> <cache id='1' level='3' type='data' size='3840' unit='KiB'/> + <monitor level='3' vcpus='1'/> </cachetune> <cachetune vcpus='4-7'> + <monitor level='3' vcpus='4-6'/> </cachetune> </cputune> In above XML, created 2 cache resctrl allocation groups and 2 resctrl monitoring groups. The changes of cache monitor will be effective in next booting of VM. 2 Show CMT result through command 'domstats' Adding the interface in qemu to report this information for resource monitor group through command 'virsh domstats --cpu-total'. Below is a typical output: # virsh domstats 1 --cpu-total Domain: 'ubuntu16.04-base' ... cpu.cache.monitor.count=2 cpu.cache.0.name=vcpus_1 cpu.cache.0.vcpus=1 cpu.cache.0.bank.count=2 cpu.cache.0.bank.0.id=0 cpu.cache.0.bank.0.bytes=4505600 cpu.cache.0.bank.1.id=1 cpu.cache.0.bank.1.bytes=5586944 cpu.cache.1.name=vcpus_4-6 cpu.cache.1.vcpus=4,5,6 cpu.cache.1.bank.count=2 cpu.cache.1.bank.0.id=0 cpu.cache.1.bank.0.bytes=17571840 cpu.cache.1.bank.1.id=1 cpu.cache.1.bank.1.bytes=29106176 Changes in v4: v4 is addressing the feature for monitoring VM vcpu thread set cache occupancy and reporting it through a virsh command. - Introduced resctrl default allocation - Introduced resctrl monitor and default monitor Changes in v3: - Addressed John Ferlan's review. - Typo fixed. - Removed VIR_ENUM_DECL(virMonitor); Changes in v2: - Introduced MBM capability. - Capability layout changed * Moved <monitor> from cahe <bank> to <cache> * Renamed <Threshold> to <reuseThreshold> - Document for 'reuseThreshold' changed. - Introduced API virResctrlInfoGetMonitorPrefix - Added more tests, covering standalone CMT, fake new feature. - Creating CMT resource control group will be subsequent job. Wang Huaqiang (18): docs: Refactor schemas to support default allocation util: Introduce resctrl monitor for CMT util: Refactor code for adding PID to the resource group util: Add interface for adding PID to monitor util: Refactor code for determining allocation path util: Add monitor interface to determine path util: Refactor code for creating resctrl group util: Add interface for creating monitor group util: Add more interfaces for resctrl monitor util: Introduce default monitor conf: Refactor code for matching existing resctrls conf: Refactor virDomainResctrlAppend conf: Add resctrl monitor configuration Util: Add function for checking if monitor is running qemu: enable resctrl monitor in qemu conf: Add a 'id' to virDomainResctrlDef qemu: refactor qemuDomainGetStatsCpu qemu: Report cache occupancy (CMT) with domstats docs/schemas/domaincommon.rng | 14 +- src/conf/domain_conf.c | 327 ++++++++++-- src/conf/domain_conf.h | 12 + src/libvirt_private.syms | 12 + src/qemu/qemu_driver.c | 272 +++++++++- src/qemu/qemu_process.c | 41 +- src/util/virresctrl.c | 555 +++++++++++++++++++-- src/util/virresctrl.h | 49 ++ tests/genericxml2xmlindata/cachetune-cdp.xml | 3 + .../cachetune-colliding-monitor.xml | 30 ++ tests/genericxml2xmlindata/cachetune-small.xml | 7 + tests/genericxml2xmltest.c | 2 + 12 files changed, 1224 insertions(+), 100 deletions(-) create mode 100644 tests/genericxml2xmlindata/cachetune-colliding-monitor.xml -- 2.7.4

1 18

[libvirt] [PATCH v2 0/8] libxl: PVHv2 support
by Marek Marczykowski-Górecki 28 Sep '18

28 Sep '18

This is a respin of my old PVHv1 patch[1], converted to PVHv2. Should the code use "PVH" name (as libxl does internally), or "PVHv2" as in many places in Xen documentation? I've chosen the former, but want to confirm it. Also, not sure about VIR_DOMAIN_OSTYPE_XENPVH (as discussed on PVHv1 patch) - while it will be messy in many cases, there is libxl_domain_build_info.u.{hvm,pv,pvh} which would be good to not mess up. Also, PVHv2 needs different kernel features than PV (CONFIG_XEN_PVH vs CONFIG_XEN_PV), so keeping the same VIR_DOMAIN_OSTYPE_XEN could be confusing. On the other hand, libxl_domain_build_info.u.pv is used in very few places (one section of libxlMakeDomBuildInfo), so guarding u.hvm access with VIR_DOMAIN_OSTYPE_HVM may be enough. For now I've reused VIR_DOMAIN_OSTYPE_XEN - in the driver itself, most of the code is the same as for PV. Since PVHv2 relies on features in newer Xen versions, I needed to convert also some older code. For example b_info->u.hvm.nested_hvm was deprecated in favor of b_info->nested_hvm. While the code do handle both old and new versions (obviously refusing PVHv2 if Xen is too old), this isn't the case for tests. How it should be handled, if at all? First few preparatory patches can be applied independently. [1] https://www.redhat.com/archives/libvir-list/2016-August/msg00376.html Changes in v2: - drop "docs: don't refer to deprecated 'linux' ostype in example" patch - migrating further away from "linux" os type is offtopic to this series and apparently is a controversial thing - drop "docs: update domain schema for machine attribute" patch - already applied - apply review comments from Jim - rebase on master Marek Marczykowski-Górecki (8): docs: add documentation of arch element of capabilities.xml libxl: set shadow memory for any guest type, not only HVM libxl: prefer new location of nested_hvm in libxl_domain_build_info libxl: reorder libxlMakeDomBuildInfo for upcoming PVH support libxl: add support for PVH tests: add basic Xen PVH test xenconfig: add support for parsing type= xl config entry xenconfig: add support for type="pvh" docs/formatcaps.html.in | 22 ++++- docs/formatdomain.html.in | 11 +- docs/schemas/domaincommon.rng | 1 +- src/libxl/libxl_capabilities.c | 33 ++++++- src/libxl/libxl_conf.c | 81 ++++++++++++----- src/libxl/libxl_driver.c | 6 +- src/xenconfig/xen_common.c | 27 +++++- src/xenconfig/xen_xl.c | 5 +- tests/libxlxml2domconfigdata/basic-pv.json | 1 +- tests/libxlxml2domconfigdata/basic-pvh.json | 49 ++++++++++- tests/libxlxml2domconfigdata/basic-pvh.xml | 28 ++++++- tests/libxlxml2domconfigdata/fullvirt-cpuid.json | 2 +- tests/libxlxml2domconfigdata/multiple-ip.json | 1 +- tests/libxlxml2domconfigdata/vnuma-hvm.json | 2 +- tests/libxlxml2domconfigtest.c | 1 +- tests/testutilsxen.c | 3 +- tests/xlconfigdata/test-fullvirt-type.cfg | 21 ++++- tests/xlconfigdata/test-fullvirt-type.xml | 27 ++++++- tests/xlconfigdata/test-paravirt-type.cfg | 13 +++- tests/xlconfigdata/test-paravirt-type.xml | 25 +++++- tests/xlconfigdata/test-pvh-type.cfg | 13 +++- tests/xlconfigdata/test-pvh-type.xml | 25 +++++- tests/xlconfigtest.c | 3 +- 23 files changed, 359 insertions(+), 41 deletions(-) create mode 100644 tests/libxlxml2domconfigdata/basic-pvh.json create mode 100644 tests/libxlxml2domconfigdata/basic-pvh.xml create mode 100644 tests/xlconfigdata/test-fullvirt-type.cfg create mode 100644 tests/xlconfigdata/test-fullvirt-type.xml create mode 100644 tests/xlconfigdata/test-paravirt-type.cfg create mode 100644 tests/xlconfigdata/test-paravirt-type.xml create mode 100644 tests/xlconfigdata/test-pvh-type.cfg create mode 100644 tests/xlconfigdata/test-pvh-type.xml base-commit: 16858439deaec0832de61c5ddb93d8e80adccf6c -- git-series 0.9.1

2 12

[libvirt] [PATCH] qemu: agent: Reset agentError when qemuConnectAgent success
by Wang Yechao 28 Sep '18

28 Sep '18

qemuAgentClose and qemuAgentIO have race condition, as follows: main thread: second thread: virEventPollRunOnce processSerialChangedEvent virEventPollDispatchHandles virMutexUnlock(&eventLoop.lock) qemuAgentClose virObjectLock(mon) virEventRemoveHandle VIR_FORCE_CLOSE(mon->fd) virObjectUnlock(mon) priv->agentError = false qemuAgentIO virObjectLock(mon) mon->fd != fd --> error = true qemuProcessHandleAgentError priv->agentError = true virObjectUnlock(mon) virMutexLock(&eventLoop.lock) qemuAgentClose set the mon->fd to '-1', and then qemuAgentIO check the mon->fd not equals to fd that registered before. qemuProcessHandleAgentError will be called to set priv->agentError to 'true', then the priv->agentError is always 'true' except restart libvirtd or restart qemu-guest-agent process in guest. We can't send any qemu-agent-command anymore even if qemuConnectAgent return success later. This is accidently occurs when hot-add-vcpu in windows2012. virsh setvcpus ... virsh qemu-agent-command $vm '{"execute":"guest-get-vcpus"}' Reset the priv->agentError to 'false' when qemuConnectAgent sucess to fix this problem. Signed-off-by: Wang Yechao <wang.yechao255(a)zte.com.cn> --- src/qemu/qemu_process.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 29b0ba1..4fbb955 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -269,6 +269,7 @@ qemuConnectAgent(virQEMUDriverPtr driver, virDomainObjPtr vm) virResetLastError(); } + priv->agentError = false; return 0; } -- 1.8.3.1

3 2

[libvirt] Entering freeze for libvirt-4.8.0
by Daniel Veillard 28 Sep '18

28 Sep '18

As planned I tagged the release in git and pushed the signed tarball and rpms to the usual place: ftp://libvirt.org/libvirt/ Seems to work for me in my limited testing, but there is some red in CI https://ci.centos.org/view/libvirt/ apparently around building. That said it worked fine for me so that could be a temporary error. The plan is to put an rc2 around Friday morning, and then if everything is fine push final verion on Oct 1st next Monday. I will keep an eye on the point raised. Please give it a try to ensure a stable and portable release ! thanks, 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/

1 1

[libvirt] [PATCH] Revert "vircgroup: cleanup controllers not managed by systemd on error"
by Pavel Hrdina 28 Sep '18

28 Sep '18

This reverts commit 1602aa28f820ada66f707cef3e536e8572fbda1e. There is no need to call virCgroupRemove() nor virCgroupFree() if virCgroupEnableMissingControllers() fails because it will not modify 'group' at all. The cleanup is done in virCgroupMakeGroup(). Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/util/vircgroup.c | 25 ++++++++++--------------- 1 file changed, 10 insertions(+), 15 deletions(-) diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c index f90906e4ad..548c873da8 100644 --- a/src/util/vircgroup.c +++ b/src/util/vircgroup.c @@ -1055,7 +1055,6 @@ virCgroupNewMachineSystemd(const char *name, int rv; virCgroupPtr init; VIR_AUTOFREE(char *) path = NULL; - virErrorPtr saved = NULL; VIR_DEBUG("Trying to setup machine '%s' via systemd", name); if ((rv = virSystemdCreateMachine(name, @@ -1088,24 +1087,20 @@ virCgroupNewMachineSystemd(const char *name, if (virCgroupEnableMissingControllers(path, pidleader, controllers, group) < 0) { - goto error; + return -1; } - if (virCgroupAddProcess(*group, pidleader) < 0) - goto error; + if (virCgroupAddProcess(*group, pidleader) < 0) { + virErrorPtr saved = virSaveLastError(); + virCgroupRemove(*group); + virCgroupFree(group); + if (saved) { + virSetError(saved); + virFreeError(saved); + } + } return 0; - - error: - saved = virSaveLastError(); - virCgroupRemove(*group); - virCgroupFree(group); - if (saved) { - virSetError(saved); - virFreeError(saved); - } - - return -1; } -- 2.17.1

4 4

[libvirt] [PATCH] vircgroup: fix NULL pointer dereferencing
by Marc Hartmayer 28 Sep '18

28 Sep '18

When virCgroupEnableMissingControllers fails it's possible that *group is still set to NULL. Therefore let's add a guard and an attribute for this. [#0] virCgroupRemove(group=0x0) [#1] virCgroupNewMachineSystemd [#2] virCgroupNewMachine [#3] qemuInitCgroup [#4] qemuSetupCgroup [#5] qemuProcessLaunch [#6] qemuProcessStart [#7] qemuDomainObjStart [#8] qemuDomainCreateWithFlags [#9] qemuDomainCreate ... Fixes: 1602aa28f820ada66f707cef3e536e8572fbda1e Reviewed-by: Boris Fiuczynski <fiuczy(a)linux.ibm.com> Reviewed-by: Bjoern Walk <bwalk(a)linux.ibm.com> Signed-off-by: Marc Hartmayer <mhartmay(a)linux.ibm.com> --- src/util/vircgroup.c | 3 ++- src/util/vircgroup.h | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c index 23957c82c7fa..06e1d158febb 100644 --- a/src/util/vircgroup.c +++ b/src/util/vircgroup.c @@ -1104,7 +1104,8 @@ virCgroupNewMachineSystemd(const char *name, error: saved = virSaveLastError(); - virCgroupRemove(*group); + if (*group) + virCgroupRemove(*group); virCgroupFree(group); if (saved) { virSetError(saved); diff --git a/src/util/vircgroup.h b/src/util/vircgroup.h index 1f676f21c380..9e1ae3706b1e 100644 --- a/src/util/vircgroup.h +++ b/src/util/vircgroup.h @@ -268,7 +268,8 @@ int virCgroupGetCpusetMemoryMigrate(virCgroupPtr group, bool *migrate); int virCgroupSetCpusetCpus(virCgroupPtr group, const char *cpus); int virCgroupGetCpusetCpus(virCgroupPtr group, char **cpus); -int virCgroupRemove(virCgroupPtr group); +int virCgroupRemove(virCgroupPtr group) + ATTRIBUTE_NONNULL(1); int virCgroupKillRecursive(virCgroupPtr group, int signum); int virCgroupKillPainfully(virCgroupPtr group); -- 2.17.0

2 2

[libvirt] [RFC PATCH v1 1/1] Add attribute multiple_mdev_support for mdev type-id
by Kirti Wankhede 28 Sep '18

28 Sep '18

Generally a single instance of mdev device, a share of physical device, is assigned to user space application or a VM. There are cases when multiple instances of mdev devices of same or different types are required by User space application or VM. For example in case of vGPU, multiple mdev devices of type which represents whole GPU can be assigned to one instance of application or VM. All types of mdev devices may not support assigning multiple mdev devices to a user space application. In that case vendor driver can fail open() call of mdev device. But there is no way to know User space application about the configuration supported by vendor driver. To expose supported configuration, vendor driver should add 'multiple_mdev_support' attribute to type-id directory if vendor driver supports assigning multiple mdev devices of a particular type-id to one instance of user space application or a VM. User space application should read if 'multiple_mdev_support' attibute is present in type-id directory of all mdev devices which are going to be used. If all read 1 then user space application can proceed with multiple mdev devices. This is optional and readonly attribute. Signed-off-by: Neo Jia <cjia(a)nvidia.com> Signed-off-by: Kirti Wankhede <kwankhede(a)nvidia.com> --- Documentation/ABI/testing/sysfs-bus-vfio-mdev | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/Documentation/ABI/testing/sysfs-bus-vfio-mdev b/Documentation/ABI/testing/sysfs-bus-vfio-mdev index 452dbe39270e..69e1291479ce 100644 --- a/Documentation/ABI/testing/sysfs-bus-vfio-mdev +++ b/Documentation/ABI/testing/sysfs-bus-vfio-mdev @@ -85,6 +85,19 @@ Users: a particular <type-id> that can help in understanding the features provided by that type of mediated device. +What: /sys/.../mdev_supported_types/<type-id>/multiple_mdev_support +Date: September 2018 +Contact: Kirti Wankhede <kwankhede(a)nvidia.com> +Description: + Reading this attribute will return 0 or 1. Returning 1 indicates + multiple mdev devices of a particular <type-id> assigned to one + User space application is supported by vendor driver. This is + optional and readonly attribute. +Users: + Userspace applications interested in knowing if multiple mdev + devices of a particular <type-id> can be assigned to one + instance of application. + What: /sys/.../<device>/<UUID>/ Date: October 2016 Contact: Kirti Wankhede <kwankhede(a)nvidia.com> -- 2.7.0

3 6

[libvirt] [PATCH 0/4] Couple of metadata locking fixes
by Michal Privoznik 27 Sep '18

27 Sep '18

Strictly speaking, only the last patch actually fixes a real problem. The first three are more of a cleanup than anything. However, I am still sending them because they make the code better. Anyway, if my explanation in 4/4 is not clear enough, here it is in simpler terms: It's important to bear in mind that at any point when connection to virtlockd is closed, virtlockd goes through its hash table and if it finds any resources acquired for the process that has closed the connection, it will kill the process. In other words, resources are per PID not per connection. For instance, if libvirtd would open two connections, acquired two different resources through each of them and then closed one it would get killed instantly because it still owns some resources. So now that this virtlockd behaviour is clear (I must say intended and desired behaviour), we can start looking into how libvirtd runs qemu. Basically, problem is that instead of opening connection to virtlockd once and keeping it open through whole run of libvirtd I used this shortcut: open the connection in virSecurityManagerMetadataLock(), dup() the connection FD, and close it in virSecurityManagerMetadataUnlock(). In theory, this works pretty well - when the duplicated FD is closed, libvirtd doesn't own any locks anymore so virtlockd doesn't kill it. What my design did not count with is fork(). On fork() the duplicated FD is cloned into the child and thus connection is closed nobody knows when. Meantime, libvirtd might continue its execution (e.g. starting another domain from another thread) and call virSecurityManagerMetadataLock(). However, before it gets to call Unlock() the child process decides to close the FD (either via exec() or exit()). Given virtlockd behaviour described couple of paragraphs above I guess it is clear to see now that virtlockd kills libvirtd. My solution to this is to always run secdriver transactions from a separate child process because on fork() only the thread that calls it is cloned. So only the thread that runs the transaction is cloned, not the one that is starting a different domain. Therefore no other fork() can occur here and therefore we are safe. I know, I know, it's complicated. But it always is around fork() and IPC. If you want to see my fix in action, just enable metadata locking, and try to start two or more domains in a loop: for ((i=0;i<1000;i++)); do \ virsh start u1604-1 & virsh start u1604-2 &\ sleep 3;\ virsh destroy u1604-1; virsh destroy u1604-2;\ done At some point you'll see virsh reporting I/O error (this is because virtlockd killed libvirtd). With my patch, I run the test multiple times without any hiccup. @Bjoern: I'm still unable to reproduce the issue you reported. However, whilst trying to do so I've came across this bug. However, my gut feeling is that this might help you. Michal Prívozník (4): security: Grab a reference to virSecurityManager for transactions virNetSocket: Be more safe with fork() around virNetSocketDupFD() virLockManagerLockDaemonAcquire: Duplicate client FD with CLOEXEC flag security: Always spawn process for transactions src/libxl/libxl_migration.c | 4 ++-- src/locking/domain_lock.c | 18 ++++++++++++++++++ src/locking/lock_driver_lockd.c | 2 +- src/qemu/qemu_migration.c | 2 +- src/rpc/virnetclient.c | 5 +++-- src/rpc/virnetclient.h | 2 +- src/rpc/virnetsocket.c | 7 ++----- src/rpc/virnetsocket.h | 2 +- src/security/security_dac.c | 17 +++++++++-------- src/security/security_selinux.c | 17 +++++++++-------- 10 files changed, 47 insertions(+), 29 deletions(-) -- 2.16.4

5 24

[libvirt] [PATCH 0/2] A few spec file fixes
by Jiri Denemark 27 Sep '18

27 Sep '18

Jiri Denemark (2): spec: Set correct TLS priority spec: Build ceph and gluster support everywhere libvirt.spec.in | 20 +++----------------- 1 file changed, 3 insertions(+), 17 deletions(-) -- 2.19.0

2 3

[libvirt] [PATCH v2 0/9] cgroup cleanups and preparation for v2
by Pavel Hrdina 27 Sep '18

27 Sep '18

Pavel Hrdina (9): vircgroup: cleanup controllers not managed by systemd on error vircgroup: fix bug in virCgroupEnableMissingControllers vircgroup: rename virCgroupAdd.*Task to virCgroupAdd.*Process vircgroup: introduce virCgroupTaskFlags vircgroup: introduce virCgroupAddThread vircgroupmock: cleanup unused cgroup files vircgroupmock: rewrite cgroup fopen mocking vircgrouptest: call virCgroupDetectMounts directly vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts src/libvirt-lxc.c | 2 +- src/libvirt_private.syms | 6 +- src/lxc/lxc_controller.c | 4 +- src/qemu/qemu_process.c | 4 +- src/qemu/qemu_tpm.c | 2 +- src/util/vircgroup.c | 143 +++++++----- src/util/vircgroup.h | 5 +- src/util/vircgrouppriv.h | 4 - tests/vircgroupdata/all-in-one.cgroups | 7 + tests/vircgroupdata/all-in-one.mounts | 2 +- tests/vircgroupdata/all-in-one.parsed | 12 +- tests/vircgroupdata/all-in-one.self.cgroup | 1 + tests/vircgroupdata/cgroups1.cgroups | 11 + tests/vircgroupdata/cgroups1.self.cgroup | 11 + tests/vircgroupdata/cgroups2.cgroups | 10 + tests/vircgroupdata/cgroups2.self.cgroup | 10 + tests/vircgroupdata/cgroups3.cgroups | 12 + tests/vircgroupdata/cgroups3.self.cgroup | 12 + tests/vircgroupdata/fedora-18.cgroups | 10 + tests/vircgroupdata/fedora-18.self.cgroup | 9 + tests/vircgroupdata/fedora-21.cgroups | 12 + tests/vircgroupdata/fedora-21.self.cgroup | 10 + tests/vircgroupdata/kubevirt.cgroups | 10 + tests/vircgroupdata/kubevirt.self.cgroup | 10 + tests/vircgroupdata/logind.cgroups | 10 + tests/vircgroupdata/logind.mounts | 2 + tests/vircgroupdata/logind.self.cgroup | 1 + tests/vircgroupdata/no-cgroups.cgroups | 8 + tests/vircgroupdata/no-cgroups.parsed | 10 - tests/vircgroupdata/no-cgroups.self.cgroup | 0 tests/vircgroupdata/ovirt-node-6.6.cgroups | 9 + .../vircgroupdata/ovirt-node-6.6.self.cgroup | 8 + tests/vircgroupdata/ovirt-node-7.1.cgroups | 11 + .../vircgroupdata/ovirt-node-7.1.self.cgroup | 10 + tests/vircgroupdata/rhel-7.1.cgroups | 11 + tests/vircgroupdata/rhel-7.1.self.cgroup | 10 + tests/vircgroupdata/systemd.cgroups | 8 + tests/vircgroupdata/systemd.mounts | 11 + tests/vircgroupdata/systemd.self.cgroup | 6 + tests/vircgroupmock.c | 206 ++---------------- tests/vircgrouptest.c | 48 ++-- 41 files changed, 399 insertions(+), 289 deletions(-) create mode 100644 tests/vircgroupdata/all-in-one.cgroups create mode 100644 tests/vircgroupdata/all-in-one.self.cgroup create mode 100644 tests/vircgroupdata/cgroups1.cgroups create mode 100644 tests/vircgroupdata/cgroups1.self.cgroup create mode 100644 tests/vircgroupdata/cgroups2.cgroups create mode 100644 tests/vircgroupdata/cgroups2.self.cgroup create mode 100644 tests/vircgroupdata/cgroups3.cgroups create mode 100644 tests/vircgroupdata/cgroups3.self.cgroup create mode 100644 tests/vircgroupdata/fedora-18.cgroups create mode 100644 tests/vircgroupdata/fedora-18.self.cgroup create mode 100644 tests/vircgroupdata/fedora-21.cgroups create mode 100644 tests/vircgroupdata/fedora-21.self.cgroup create mode 100644 tests/vircgroupdata/kubevirt.cgroups create mode 100644 tests/vircgroupdata/kubevirt.self.cgroup create mode 100644 tests/vircgroupdata/logind.cgroups create mode 100644 tests/vircgroupdata/logind.mounts create mode 100644 tests/vircgroupdata/logind.self.cgroup create mode 100644 tests/vircgroupdata/no-cgroups.cgroups delete mode 100644 tests/vircgroupdata/no-cgroups.parsed create mode 100644 tests/vircgroupdata/no-cgroups.self.cgroup create mode 100644 tests/vircgroupdata/ovirt-node-6.6.cgroups create mode 100644 tests/vircgroupdata/ovirt-node-6.6.self.cgroup create mode 100644 tests/vircgroupdata/ovirt-node-7.1.cgroups create mode 100644 tests/vircgroupdata/ovirt-node-7.1.self.cgroup create mode 100644 tests/vircgroupdata/rhel-7.1.cgroups create mode 100644 tests/vircgroupdata/rhel-7.1.self.cgroup create mode 100644 tests/vircgroupdata/systemd.cgroups create mode 100644 tests/vircgroupdata/systemd.mounts create mode 100644 tests/vircgroupdata/systemd.self.cgroup -- 2.17.1

4 23

[libvirt] [PATCH for 4.8.0] qemu: Temporarily disable metadata locking
by Michal Privoznik 27 Sep '18

27 Sep '18

Turns out, there are couple of bugs that prevent this feature from being operational. Given how close to the release we are disable the feature temporarily. Hopefully, it can be enabled back after all the bugs are fixed. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- There are two major problems: 1) when using regular locking the lockspace code gets confused and starting domain is denied after hitting 60 second timeout. 2) the virtlockd connection FD might leak into children process resulting in sudden killing of libvirtd. I posted a patch for 2), and I'm working on patch for 1). src/qemu/libvirtd_qemu.aug | 1 - src/qemu/qemu.conf | 9 --------- src/qemu/qemu_conf.c | 11 ----------- src/qemu/test_libvirtd_qemu.aug.in | 1 - 4 files changed, 22 deletions(-) diff --git a/src/qemu/libvirtd_qemu.aug b/src/qemu/libvirtd_qemu.aug index 42e325d4fb..ddc4bbfd1d 100644 --- a/src/qemu/libvirtd_qemu.aug +++ b/src/qemu/libvirtd_qemu.aug @@ -98,7 +98,6 @@ module Libvirtd_qemu = | bool_entry "relaxed_acs_check" | bool_entry "allow_disk_format_probing" | str_entry "lock_manager" - | str_entry "metadata_lock_manager" let rpc_entry = int_entry "max_queued" | int_entry "keepalive_interval" diff --git a/src/qemu/qemu.conf b/src/qemu/qemu.conf index 84492719c4..8391332cb4 100644 --- a/src/qemu/qemu.conf +++ b/src/qemu/qemu.conf @@ -659,15 +659,6 @@ #lock_manager = "lockd" -# To serialize two or more daemons trying to change metadata on a -# file (e.g. a file on NFS share), libvirt offers a locking -# mechanism. Currently, only "lockd" is supported (or no locking -# at all if unset). Note that this is independent of lock_manager -# described above. -# -#metadata_lock_manager = "lockd" - - # Set limit of maximum APIs queued on one domain. All other APIs # over this threshold will fail on acquiring job lock. Specially, # setting to zero turns this feature off. diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 33508174cb..fc84186a7e 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -838,17 +838,6 @@ int virQEMUDriverConfigLoadFile(virQEMUDriverConfigPtr cfg, if (virConfGetValueString(conf, "lock_manager", &cfg->lockManagerName) < 0) goto cleanup; - if (virConfGetValueString(conf, "metadata_lock_manager", - &cfg->metadataLockManagerName) < 0) - goto cleanup; - if (cfg->metadataLockManagerName && - STRNEQ(cfg->metadataLockManagerName, "lockd")) { - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("unknown metadata lock manager name %s"), - cfg->metadataLockManagerName); - goto cleanup; - } - if (virConfGetValueString(conf, "stdio_handler", &stdioHandler) < 0) goto cleanup; if (stdioHandler) { diff --git a/src/qemu/test_libvirtd_qemu.aug.in b/src/qemu/test_libvirtd_qemu.aug.in index 451e73126e..f1e8806ad2 100644 --- a/src/qemu/test_libvirtd_qemu.aug.in +++ b/src/qemu/test_libvirtd_qemu.aug.in @@ -81,7 +81,6 @@ module Test_libvirtd_qemu = { "mac_filter" = "1" } { "relaxed_acs_check" = "1" } { "lock_manager" = "lockd" } -{ "metadata_lock_manager" = "lockd" } { "max_queued" = "0" } { "keepalive_interval" = "5" } { "keepalive_count" = "5" } -- 2.16.4

2 1

[libvirt] [PATCH v4 00/23] Introduce metadata locking
by Michal Privoznik 27 Sep '18

27 Sep '18

Technically, this is v4 of: https://www.redhat.com/archives/libvir-list/2018-August/msg01627.html However, this is implementing different approach than any of the previous versions. One of the problems with previous version was that it was too complicated. The main reason for that was that we could not close the connection whilst there was a file locked. So we had to invent a mechanism that would prevent that (on the client side). These patches implement different approach. They rely on secdriver's transactions which bring all the paths we want to label into one place so that they can be relabelled within different namespace. I'm extending this idea so that transactions run all the time (regardless of domain namespacing) and only at the very last moment is decided which namespace would the relabeling run in. Metadata locking is then as easy as putting lock/unlock calls around one function. You can find the patches at my github too: https://github.com/zippy2/libvirt/tree/disk_metadata_lock_v4_alt Michal Prívozník (23): qemu_security: Fully implement qemuSecurityDomainSetPathLabel qemu_security: Fully implement qemuSecurity{Set,Restore}SavedStateLabel qemu_security: Require full wrappers for APIs that might touch a file virSecurityManagerTransactionCommit: Accept pid == -1 qemu_security: Run transactions more frequently virlockspace: Allow caller to specify start and length offset in virLockSpaceAcquireResource lock_driver_lockd: Introduce VIR_LOCK_SPACE_PROTOCOL_ACQUIRE_RESOURCE_METADATA flag lock_driver: Introduce new VIR_LOCK_MANAGER_OBJECT_TYPE_DAEMON _virLockManagerLockDaemonPrivate: Move @hasRWDisks into dom union lock_driver: Introduce VIR_LOCK_MANAGER_RESOURCE_TYPE_METADATA lock_driver: Introduce VIR_LOCK_MANAGER_ACQUIRE_ROLLBACK lock_daemon_dispatch: Check for ownerPid rather than ownerId lock_manager: Allow disabling configFile for virLockManagerPluginNew qemu_conf: Introduce metadata_lock_manager security_manager: Load lock plugin on init security_manager: Introduce metadata locking APIs security_dac: Move transaction handling up one level security_dac: Fix info messages when chown()-ing security_dac: Lock metadata when running transaction virSecuritySELinuxRestoreFileLabel: Rename 'err' label virSecuritySELinuxRestoreFileLabel: Adjust code pattern security_selinux: Move transaction handling up one level security_dac: Lock metadata when running transaction cfg.mk | 4 +- src/locking/lock_daemon_dispatch.c | 25 ++- src/locking/lock_driver.h | 12 ++ src/locking/lock_driver_lockd.c | 417 +++++++++++++++++++++++++------------ src/locking/lock_driver_lockd.h | 1 + src/locking/lock_driver_sanlock.c | 44 ++-- src/locking/lock_manager.c | 10 +- src/lxc/lxc_controller.c | 3 +- src/lxc/lxc_driver.c | 2 +- src/qemu/libvirtd_qemu.aug | 1 + src/qemu/qemu.conf | 8 + src/qemu/qemu_conf.c | 13 ++ src/qemu/qemu_conf.h | 1 + src/qemu/qemu_domain.c | 3 +- src/qemu/qemu_driver.c | 10 +- src/qemu/qemu_process.c | 15 +- src/qemu/qemu_security.c | 272 +++++++++++++++++------- src/qemu/qemu_security.h | 18 +- src/qemu/test_libvirtd_qemu.aug.in | 1 + src/security/security_dac.c | 134 ++++++++---- src/security/security_manager.c | 171 ++++++++++++++- src/security/security_manager.h | 9 + src/security/security_selinux.c | 118 ++++++++--- src/util/virlockspace.c | 15 +- src/util/virlockspace.h | 4 + tests/seclabeltest.c | 2 +- tests/securityselinuxlabeltest.c | 2 +- tests/securityselinuxtest.c | 2 +- tests/testutilsqemu.c | 2 +- tests/virlockspacetest.c | 29 ++- 30 files changed, 1006 insertions(+), 342 deletions(-) -- 2.16.4

4 70

[libvirt] [PATCH v3 0/3] fix build failure in vircgroup code
by Pavel Hrdina 27 Sep '18

27 Sep '18

Chagnes in v3: - removed VIR_CGROUP_SUPPORTED - include system headers only on linux Pavel Hrdina (3): vircgroup: remove VIR_CGROUP_SUPPORTED vircgroup: include system headers only on linux vircgroupv1: fix build on non-linux OSes src/util/vircgroup.c | 38 ++++++++++++++++---------------------- src/util/vircgroupv1.c | 20 +++++++++++++++----- 2 files changed, 31 insertions(+), 27 deletions(-) -- 2.17.1

2 4

[libvirt] [tck PATCH] Fix typo in StrictHostKeyChecking setting
by Laine Stump 26 Sep '18

26 Sep '18

This was inadvertantly changed from "no" to "off" in commit 685bf11b. Signed-off-by: Laine Stump <laine(a)laine.org> --- ... and noticed soon after, but apparently nobody took the time to push a fix. Pushed under the trivial rule. scripts/nwfilter/210-no-mac-spoofing.t | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/nwfilter/210-no-mac-spoofing.t b/scripts/nwfilter/210-no-mac-spoofing.t index 87c19e7..78c500c 100644 --- a/scripts/nwfilter/210-no-mac-spoofing.t +++ b/scripts/nwfilter/210-no-mac-spoofing.t @@ -97,7 +97,7 @@ my $ssh = Net::OpenSSH->new($guestip, user => "root", password => $tck->root_password(), master_opts => [-o => "UserKnownHostsFile=/dev/null", - -o => "StrictHostKeyChecking=off"]); + -o => "StrictHostKeyChecking=no"]); # now bring eth0 down, change MAC and bring it up again diag "fiddling with mac"; -- 2.17.1

1 0

[libvirt] [PATCH] tests: libxl: skip tests with nested_hvm
by Ján Tomko 26 Sep '18

26 Sep '18

Commit 95d19cd unconditionally adjusted the tests to account for the conditional move of the nested_hvm setting location. Run the affected tests only for the new setup (witnessed by LIBXL_HAVE_BUILDINFO_NESTED_HVM). Signed-off-by: Ján Tomko <jtomko(a)redhat.com> --- Technically a build breaker fix, but I'd love to hear any ideas that would easily let us test both versions. Also, my app-emulation/xen-tools-4.9.2-r1 on Gentoo does not seem to have LIBXL_HAVE_BUILDINFO_NESTED_HVM in libxl.h, even though it was introduced in 4.9.0-rc7-831-g3c44f8ed44 tests/libxlxml2domconfigtest.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/tests/libxlxml2domconfigtest.c b/tests/libxlxml2domconfigtest.c index 0eba5814b1..6b1f9826d0 100644 --- a/tests/libxlxml2domconfigtest.c +++ b/tests/libxlxml2domconfigtest.c @@ -207,9 +207,16 @@ mymain(void) DO_TEST("cpu-shares-hvm"); DO_TEST("variable-clock-hvm"); DO_TEST("moredevs-hvm"); + +# ifdef LIBXL_HAVE_BUILDINFO_NESTED_HVM DO_TEST("vnuma-hvm"); +# endif + DO_TEST("multiple-ip"); + +# ifdef LIBXL_HAVE_BUILDINFO_NESTED_HVM DO_TEST("fullvirt-cpuid"); +# endif unlink("libxl-driver.log"); -- 2.16.4

3 3

[libvirt] [PATCH v2] vircgroupv1: add ifdef around cgroup code
by Pavel Hrdina 26 Sep '18

26 Sep '18

Cgroups are supported only on linux. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- changed in v2: - define moved to vircgrouppriv.h - fixed alternative virCgroupV1Register function src/util/vircgroup.c | 5 ----- src/util/vircgrouppriv.h | 8 ++++++++ src/util/vircgroupv1.c | 13 +++++++++++++ 3 files changed, 21 insertions(+), 5 deletions(-) diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c index 23957c82c7..aae699db0e 100644 --- a/src/util/vircgroup.c +++ b/src/util/vircgroup.c @@ -66,11 +66,6 @@ VIR_LOG_INIT("util.cgroup"); #define CGROUP_NB_TOTAL_CPU_STAT_PARAM 3 #define CGROUP_NB_PER_CPU_STAT_PARAM 1 -#if defined(__linux__) && defined(HAVE_GETMNTENT_R) && \ - defined(_DIRENT_HAVE_D_TYPE) && defined(_SC_CLK_TCK) -# define VIR_CGROUP_SUPPORTED -#endif - VIR_ENUM_IMPL(virCgroupController, VIR_CGROUP_CONTROLLER_LAST, "cpu", "cpuacct", "cpuset", "memory", "devices", "freezer", "blkio", "net_cls", "perf_event", diff --git a/src/util/vircgrouppriv.h b/src/util/vircgrouppriv.h index c50a25f195..5a19f965c2 100644 --- a/src/util/vircgrouppriv.h +++ b/src/util/vircgrouppriv.h @@ -29,9 +29,17 @@ #ifndef __VIR_CGROUP_PRIV_H__ # define __VIR_CGROUP_PRIV_H__ +# include <dirent.h> +# include <unistd.h> + # include "vircgroup.h" # include "vircgroupbackend.h" +# if defined(__linux__) && defined(HAVE_GETMNTENT_R) && \ + defined(_DIRENT_HAVE_D_TYPE) && defined(_SC_CLK_TCK) +# define VIR_CGROUP_SUPPORTED +# endif + struct _virCgroupV1Controller { int type; char *mountPoint; diff --git a/src/util/vircgroupv1.c b/src/util/vircgroupv1.c index 62a6e5c448..a00dd06148 100644 --- a/src/util/vircgroupv1.c +++ b/src/util/vircgroupv1.c @@ -55,6 +55,8 @@ VIR_ENUM_IMPL(virCgroupV1Controller, VIR_CGROUP_CONTROLLER_LAST, "name=systemd"); +#ifdef VIR_CGROUP_SUPPORTED + /* We're looking for at least one 'cgroup' fs mount, * which is *not* a named mount. */ static bool @@ -2099,3 +2101,14 @@ virCgroupV1Register(void) { virCgroupBackendRegister(&virCgroupV1Backend); } + +#else /* !VIR_CGROUP_SUPPORTED */ + +void +virCgroupV1Register(void) +{ + virReportSystemError(ENOSYS, "%s", + _("Control groups not supported on this platform")); +} + +#endif /* !VIR_CGROUP_SUPPORTED */ -- 2.17.1

2 1

[libvirt] [PATCH v3] qemu: Remove network type limitation for qemuARPGetInterfaces
by Lin Ma 26 Sep '18

26 Sep '18

Let's ignore the checking of interface type when we call the function qemuARPGetInterfaces to get IP from host's arp table. Based on suggestion from Laine. https://www.redhat.com/archives/libvir-list/2018-September/msg00684.html Signed-off-by: Lin Ma <lma(a)suse.com> --- src/qemu/qemu_driver.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 10d6bca186..3110e74e0e 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -20945,9 +20945,6 @@ qemuARPGetInterfaces(virDomainObjPtr vm, goto cleanup; for (i = 0; i < vm->def->nnets; i++) { - if (vm->def->nets[i]->type != VIR_DOMAIN_NET_TYPE_NETWORK) - continue; - virMacAddrFormat(&(vm->def->nets[i]->mac), macaddr); for (j = 0; j < table->n; j++) { virArpTableEntry entry = table->t[j]; -- 2.19.0

3 2

[libvirt] [PATCH] virsh: Require explicit --domain for domxml-to-native
by Jiri Denemark 26 Sep '18

26 Sep '18

The domxml-to-native virsh command accepts either --xml or --domain option followed by a file or domain name respectively. The --domain option is documented as required, which means an argument with no option is treated as --xml. Commit v4.3.0-127-gd86531daf2 broke this by making --domain optional and thus an argument with no option was treated as --domain. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- tools/virsh-domain.c | 6 +++--- tools/virsh.h | 8 +++++--- 2 files changed, 8 insertions(+), 6 deletions(-) diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c index cfb0095333..372bdb95d3 100644 --- a/tools/virsh-domain.c +++ b/tools/virsh-domain.c @@ -9509,7 +9509,7 @@ static const vshCmdInfo info_qemu_monitor_event[] = { static const vshCmdOptDef opts_qemu_monitor_event[] = { VIRSH_COMMON_OPT_DOMAIN_OT_STRING(N_("filter by domain name, id or uuid"), - 0), + 0, 0), {.name = "event", .type = VSH_OT_STRING, .help = N_("filter by event name") @@ -10062,7 +10062,7 @@ static const vshCmdOptDef opts_domxmltonative[] = { .flags = VSH_OFLAG_REQ, .help = N_("target config data type format") }, - VIRSH_COMMON_OPT_DOMAIN_OT_STRING_FULL(0), + VIRSH_COMMON_OPT_DOMAIN_OT_STRING_FULL(VSH_OFLAG_REQ_OPT, 0), {.name = "xml", .type = VSH_OT_STRING, .help = N_("xml data file to export from") @@ -13312,7 +13312,7 @@ static const vshCmdInfo info_event[] = { static const vshCmdOptDef opts_event[] = { VIRSH_COMMON_OPT_DOMAIN_OT_STRING(N_("filter by domain name, id or uuid"), - 0), + 0, 0), {.name = "event", .type = VSH_OT_STRING, .completer = virshDomainEventNameCompleter, diff --git a/tools/virsh.h b/tools/virsh.h index af0d6ce686..6ff5c74b5e 100644 --- a/tools/virsh.h +++ b/tools/virsh.h @@ -107,16 +107,18 @@ .help = _helpstr \ } -# define VIRSH_COMMON_OPT_DOMAIN_OT_STRING(_helpstr, cflags) \ +# define VIRSH_COMMON_OPT_DOMAIN_OT_STRING(_helpstr, oflags, cflags) \ {.name = "domain", \ .type = VSH_OT_STRING, \ + .flags = oflags, \ .help = _helpstr, \ .completer = virshDomainNameCompleter, \ .completer_flags = cflags, \ } -# define VIRSH_COMMON_OPT_DOMAIN_OT_STRING_FULL(cflags) \ - VIRSH_COMMON_OPT_DOMAIN_OT_STRING(N_("domain name, id or uuid"), cflags) +# define VIRSH_COMMON_OPT_DOMAIN_OT_STRING_FULL(oflags, cflags) \ + VIRSH_COMMON_OPT_DOMAIN_OT_STRING(N_("domain name, id or uuid"), \ + oflags, cflags) # define VIRSH_COMMON_OPT_DOMAIN_OT_ARGV(_helpstr, cflags) \ {.name = "domain", \ -- 2.19.0

2 2

[libvirt] Plans for next release
by Daniel Veillard 26 Sep '18

26 Sep '18

that's next week ! If we want to release around Oct 1st, I would suggest to enter freeze this Wed, probably in the morning europe time, then plan for an RC2 friday morning and if everything goes well, have the final release next monday. I hope that works for everybody ! 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 7

[libvirt] [PATCH 0/7] configure: remove some unused checks
by Ján Tomko 26 Sep '18

26 Sep '18

Ján Tomko (7): configure: split common header and function checks configure: sort AC_CHECK_HEADERS argument list configure: remove duplicit check for sys/un.h configure: remove check for regex.h configure: remove check for poll.h configure: do not check for kill configure: remove regexec and getgrnam_r checks configure.ac | 50 +++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 41 insertions(+), 9 deletions(-) -- 2.16.4

2 14

[libvirt] [PATCH 0/4] Small changes to virdbus
by Marc Hartmayer 26 Sep '18

26 Sep '18

Marc Hartmayer (4): virdbus: Grab a ref as long as the while loop is executed virdbus: Unref the D-Bus connection when closing virdbus: Report a debug message that dbus_watch_handle() has failed virdbus: Use the mnemonic macros for dbus_bool_t values src/util/virdbus.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) -- 2.17.0

2 6

[libvirt] [PATCH] lxc_monitor: Avoid AB / BA lock race
by Mark Asselstine 25 Sep '18

25 Sep '18

A deadlock situation can occur when autostarting a LXC domain 'guest' due to two threads attempting to take opposing locks while holding opposing locks (AB BA problem). Thread A takes and holds the 'vm' lock while attempting to take the 'client' lock, meanwhile, thread B takes and holds the 'client' lock while attempting to take the 'vm' lock. The potential for this can be seen as follows: Thread A: virLXCProcessAutostartDomain (takes vm lock) --> virLXCProcessStart --> virLXCProcessConnectMonitor --> virLXCMonitorNew --> virNetClientSetCloseCallback (wants client lock) Thread B: virNetClientIncomingEvent (takes client lock) --> virNetClientIOHandleInput --> virNetClientCallDispatch --> virNetClientCallDispatchMessage --> virNetClientProgramDispatch --> virLXCMonitorHandleEventInit --> virLXCProcessMonitorInitNotify (wants vm lock) Since these threads are scheduled independently and are preemptible it is possible for the deadlock scenario to occur where each thread locks their first lock but both will fail to get their second lock and just spin forever. You get something like: virLXCProcessAutostartDomain (takes vm lock) --> virLXCProcessStart --> virLXCProcessConnectMonitor --> virLXCMonitorNew <...> virNetClientIncomingEvent (takes client lock) --> virNetClientIOHandleInput --> virNetClientCallDispatch --> virNetClientCallDispatchMessage --> virNetClientProgramDispatch --> virLXCMonitorHandleEventInit --> virLXCProcessMonitorInitNotify (wants vm lock but spins) <...> --> virNetClientSetCloseCallback (wants client lock but spins) Neither thread ever gets the lock it needs to be able to continue while holding the lock that the other thread needs. The actual window for preemption which can cause this deadlock is rather small, between the calls to virNetClientProgramNew() and execution of virNetClientSetCloseCallback(), both in virLXCMonitorNew(). But it can be seen in real world use that this small window is enough. By moving the call to virNetClientSetCloseCallback() ahead of virNetClientProgramNew() we can close any possible chance of the deadlock taking place. There should be no other implications to the move since the close callback (in the unlikely event was called) will spin on the vm lock. The remaining work that takes place between the old call location of virNetClientSetCloseCallback() and the new location is unaffected by the move. Signed-off-by: Mark Asselstine <mark.asselstine(a)windriver.com> --- NOTE I am not intimately familiar with the libvirt code, so although I am confident of the cause of the issue I am not sure this is the best solution. So I am really curious what people who do know this code think and as such I am definitely open to discussions and suggestions on how to rework things to properly address the issue. src/lxc/lxc_monitor.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/src/lxc/lxc_monitor.c b/src/lxc/lxc_monitor.c index e765c16..218f426 100644 --- a/src/lxc/lxc_monitor.c +++ b/src/lxc/lxc_monitor.c @@ -161,6 +161,10 @@ virLXCMonitorPtr virLXCMonitorNew(virDomainObjPtr vm, if (virNetClientRegisterAsyncIO(mon->client) < 0) goto error; + /* avoid deadlock by making this call before assigning virLXCMonitorEvents */ + virNetClientSetCloseCallback(mon->client, virLXCMonitorEOFNotify, mon, + virLXCMonitorCloseFreeCallback); + if (!(mon->program = virNetClientProgramNew(VIR_LXC_MONITOR_PROGRAM, VIR_LXC_MONITOR_PROGRAM_VERSION, virLXCMonitorEvents, @@ -176,8 +180,6 @@ virLXCMonitorPtr virLXCMonitorNew(virDomainObjPtr vm, memcpy(&mon->cb, cb, sizeof(mon->cb)); virObjectRef(mon); - virNetClientSetCloseCallback(mon->client, virLXCMonitorEOFNotify, mon, - virLXCMonitorCloseFreeCallback); cleanup: VIR_FREE(sockpath); -- 2.7.4

2 3

[libvirt] [PATCH v2 0/5] qemu: Change the way we generate VIR_DOMAIN_EVENT_RESUMED
by Jiri Denemark 25 Sep '18

25 Sep '18

https://bugzilla.redhat.com/show_bug.cgi?id=1612943 Version 2: - keep VIR_DOMAIN_EVENT_RESUMED_MIGRATED event at the end of post-copy migration Jiri Denemark (5): qemu: Properly report VIR_DOMAIN_EVENT_RESUMED_FROM_SNAPSHOT qemu: Report more appropriate running reasons qemu: Pass running reason to RESUME event handler qemu: Map running reason to resume event detail qemu: Avoid duplicate resume events and state changes src/qemu/qemu_domain.c | 29 +++++++++++++++++++++++ src/qemu/qemu_domain.h | 7 ++++++ src/qemu/qemu_driver.c | 13 ----------- src/qemu/qemu_migration.c | 49 ++++++++++++++++----------------------- src/qemu/qemu_process.c | 46 ++++++++++++++++++++++-------------- 5 files changed, 85 insertions(+), 59 deletions(-) -- 2.19.0

2 6

[libvirt] [PATCH] vircgroupv1: add ifdef around cgroup code
by Pavel Hrdina 25 Sep '18

25 Sep '18

Cgroups are supported only on linux. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/util/vircgroupv1.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/src/util/vircgroupv1.c b/src/util/vircgroupv1.c index 62a6e5c448..3d05fbd745 100644 --- a/src/util/vircgroupv1.c +++ b/src/util/vircgroupv1.c @@ -55,6 +55,8 @@ VIR_ENUM_IMPL(virCgroupV1Controller, VIR_CGROUP_CONTROLLER_LAST, "name=systemd"); +#ifdef VIR_CGROUP_SUPPORTED + /* We're looking for at least one 'cgroup' fs mount, * which is *not* a named mount. */ static bool @@ -2099,3 +2101,15 @@ virCgroupV1Register(void) { virCgroupBackendRegister(&virCgroupV1Backend); } + +#else /* !VIR_CGROUP_SUPPORTED */ + +void +virCgroupV1Register(void) +{ + virReportSystemError(ENOSYS, "%s", + _("Control groups not supported on this platform")); + return -1; +} + +#endif /* !VIR_CGROUP_SUPPORTED */ -- 2.17.1

3 2