April 2019 - Devel - libvirt List Archives

[libvirt] RFC: Introducing Memory Bandwidth Monitoring (MBM)
by Huaqiang,Wang 26 Apr '19

26 Apr '19

RFC: Introducing Memory Bandwidth Monitoring (MBM) ================================================== Kernel has removed the interfaces of getting memory bandwidth utilization interfaces from `perf` sub-module, refer to https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id…. Afterwards, kernel implements the `resctrl` sub-module and provides the alternative interfaces to allocate/monitor resource of cache/memory bandwidth. The cache monitoring based on `resctr` interface is already implemented, this RFC is to raise discussion on implement the memory bandwidth monitoring functionality. Very similar to the way of implementing 'cache monitoring', to support memory bandwidth monitoring, it could be considered in three aspects: * Identifying the host capability for MBM * Creating MBM groups * Reporting the result of the monitoring Identify the host capability of MBM ----------------------------------- Host MBM feature coud be identified through the standard way of enumerating host capability, such as command `virsh capability <domain>`. A host with MBM, the result of command above may contain following lines:: <capabilities> <host> … <cache> <bank id='0' level='3' type='both' size='15' unit='MiB' cpus='0-5'> <control granularity='768' min='1536' unit='KiB' type='both' maxAllocs='4'/> </bank> <monitor level='3' reuseThreshold='270336' maxMonitors='176'> <feature name='llc_occupancy'/> </monitor> </cache> <memory_bandwidth> <node id='0' cpus='0-5'> <control granularity='10' min ='10' maxAllocs='4'/> </node> <node id='1' cpus='6-11'> <control granularity='10' min ='10' maxAllocs='4'/> </node> + <monitor maxMonitors='176'> + <feature name='mbm_total_bytes'/> + <feature name='mbm_local_bytes'/> + </monitor> </memory_bandwidth> </host> </capabilities> The <monitor> element under <memory_bandwidth> indicates a memory bandwidth monitor exists, and the property 'maxMonitors' tells how many monitors could be generated, which is shared with cache monitor. The feature list inside the <monitor> element indicates MBM feature set. Create the MBM group -------------------- It is supported to create multiple monitors to a vm instance and each monitor is capable to track the memory bandwidth utilization accumulatively in bytes for specific vCPUs. To create a memory bandwidth monitor for specific vCPUs, for example create 2 memory bandwidth monitors, one is tracking the memory BW usage for vCPU0 and another is for vCPU 0-4, then you can set the request for two monitors in following way by editing the domain XML file:: <cputune> <memorytune vcpus='0-4'> <node id='0' bandwidth='20'/> <node id='1' bandwidth='30'/> + <monitor vcpus='0-4'/> + <monitor vcpus='0'/> </memorytune> </cputune> Report the result of monitor ---------------------------- The statistical results of memory bandwidth monitors could be demonstrated through command `virsh domstat <doman>` in following arrangement. virsh domstats <domain>:: ... memory.bw.monitor.count=<total number of MBM monitors> memory.bw.monitor.<monitor index>.name=<resctrl group name for this monitor> memory.bw.monitor.<monitor index>.vcpus=<vcpu set list monitored by this monitor> memory.bw.monitor.<monitor index>.controller.count=<memory controller number> memory.bw.monitor.<monitor index>.controller.<index>.local.bytes=<bytes> memory.bw.monitor.<monitor index>.controller.<index>.total.bytes=<bytes> ...

1 0

[libvirt] [PATCH] Handle copying bitmaps to larger data buffers
by Allen, John 25 Apr '19

25 Apr '19

If a bitmap of a shorter length than the data buffer is passed to virBitmapToDataBuf, it will read off the end of the bitmap and copy junk into the returned buffer. Add a check to only copy the length of the bitmap to the buffer. The problem can be observed after setting a vcpu affinity using the vcpupin command on a system with a large number of cores: # virsh vcpupin example_domain 0 0 # virsh vcpupin example_domain 0 VCPU CPU Affinity --------------------------- 0 0,192,197-198,202 Signed-off-by: John Allen <john.allen(a)amd.com> --- diff --git a/src/util/virbitmap.c b/src/util/virbitmap.c index d074f29e54..021174cbb3 100644 --- a/src/util/virbitmap.c +++ b/src/util/virbitmap.c @@ -824,11 +824,15 @@ virBitmapToDataBuf(virBitmapPtr bitmap, unsigned char *bytes, size_t len) { + size_t nbytes = bitmap->map_len * (VIR_BITMAP_BITS_PER_UNIT / CHAR_BIT); unsigned long *l; size_t i, j; memset(bytes, 0, len); + /* If bitmap and buffer differ in size, only fill to the smaller length */ + len = MIN(len, nbytes); + /* htole64 is not provided by gnulib, so we do the conversion by hand */ l = bitmap->map; for (i = j = 0; i < len; i++, j++) {

3 3

[libvirt] [PATCH] cpu_map: Add support for cldemote CPU feature
by Jiri Denemark 25 Apr '19

25 Apr '19

Added in QEMU by v2.12.0-481-g0da0fb0628 (released in 3.0). Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/cpu_map/x86_features.xml | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/cpu_map/x86_features.xml b/src/cpu_map/x86_features.xml index 02431bea29..efcc10b1ae 100644 --- a/src/cpu_map/x86_features.xml +++ b/src/cpu_map/x86_features.xml @@ -310,6 +310,9 @@ <feature name='la57'> <cpuid eax_in='0x07' ecx_in='0x00' ecx='0x00010000'/> </feature> + <feature name='cldemote'> + <cpuid eax_in='0x07' ecx_in='0x00' ecx='0x02000000'/> + </feature> <feature name='avx512-4vnniw'> <cpuid eax_in='0x07' ecx_in='0x00' edx='0x00000004'/> -- 2.21.0

2 1

[libvirt] [PATCH v4 0/3] Don't pm suspend guest if it's unable to wake up
by Daniel Henrique Barboza 25 Apr '19

25 Apr '19

changes from v3: - patch 3 removed. We no longer consider PM_WAKEUP_SUPPORT a cap. It is confusing to have a 'runtime cap' that can't be resolved inside qemu_capabilities.c. The suspend operation isn't time critical per se, and a call to QMP each time 'virsh dompmsuspend' is executed is not dreadful to the user experience. More information and discussion can be found at [1]. - due to the change above, patch 4 now executes the query-current-machine QMP call (if the running QEMU instance supports it) during qemuDomainPMSuspendForDuration to check if the guest has wakeup support. - previous patch link: [1] https://www.redhat.com/archives/libvir-list/2019-April/msg00767.html Daniel Henrique Barboza (2): qemu_capabilities: Add QEMU_CAPS_QUERY_CURRENT_MACHINE qemuDomainPMSuspendForDuration: check for wake-up support Michal Prívozník (1): qemu_monitor: Introduce handler for 'query-current-machine' command src/qemu/qemu_capabilities.c | 2 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_driver.c | 54 +++++++++++++++++++ src/qemu/qemu_monitor.c | 10 ++++ src/qemu/qemu_monitor.h | 9 ++++ src/qemu/qemu_monitor_json.c | 50 +++++++++++++++++ src/qemu/qemu_monitor_json.h | 5 ++ .../caps_4.0.0.riscv32.xml | 1 + .../caps_4.0.0.riscv64.xml | 1 + .../caps_4.0.0.x86_64.xml | 1 + 10 files changed, 134 insertions(+) -- 2.20.1

2 7

[libvirt] [PATCH] vircgroup: no need to ifdef virCgroupFree
by Pavel Hrdina 25 Apr '19

25 Apr '19

virCgroup struct is always defined and the free function is not calling anything that would require OS supporting cgroups. This fixes an issue if we try to start a VM with QEMU binary that doesn't support QXL. The start operation will fail in qemuProcessStartValidateVideo() which will set correct error message, but later in one of the cleanup paths we will call qemuDomainObjPrivateDataClear() which always calls virCgroupFree() and that will fail on OS that doesn't support cgroups and it will set a new error which will be eventually reported to user. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/util/vircgroup.c | 62 +++++++++++++++++++------------------------- 1 file changed, 27 insertions(+), 35 deletions(-) diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c index 4238d7014b..f58e336404 100644 --- a/src/util/vircgroup.c +++ b/src/util/vircgroup.c @@ -1279,33 +1279,6 @@ virCgroupNewIgnoreError(void) } -/** - * virCgroupFree: - * - * @group: The group structure to free - */ -void -virCgroupFree(virCgroupPtr *group) -{ - size_t i; - - if (*group == NULL) - return; - - for (i = 0; i < VIR_CGROUP_CONTROLLER_LAST; i++) { - VIR_FREE((*group)->legacy[i].mountPoint); - VIR_FREE((*group)->legacy[i].linkPoint); - VIR_FREE((*group)->legacy[i].placement); - } - - VIR_FREE((*group)->unified.mountPoint); - VIR_FREE((*group)->unified.placement); - - VIR_FREE((*group)->path); - VIR_FREE(*group); -} - - /** * virCgroupHasController: query whether a cgroup controller is present * @@ -2917,14 +2890,6 @@ virCgroupNewIgnoreError(void) } -void -virCgroupFree(virCgroupPtr *group ATTRIBUTE_UNUSED) -{ - virReportSystemError(ENXIO, "%s", - _("Control groups not supported on this platform")); -} - - bool virCgroupHasController(virCgroupPtr cgroup ATTRIBUTE_UNUSED, int controller ATTRIBUTE_UNUSED) @@ -3565,6 +3530,33 @@ virCgroupControllerAvailable(int controller ATTRIBUTE_UNUSED) #endif /* !__linux__ */ +/** + * virCgroupFree: + * + * @group: The group structure to free + */ +void +virCgroupFree(virCgroupPtr *group) +{ + size_t i; + + if (*group == NULL) + return; + + for (i = 0; i < VIR_CGROUP_CONTROLLER_LAST; i++) { + VIR_FREE((*group)->legacy[i].mountPoint); + VIR_FREE((*group)->legacy[i].linkPoint); + VIR_FREE((*group)->legacy[i].placement); + } + + VIR_FREE((*group)->unified.mountPoint); + VIR_FREE((*group)->unified.placement); + + VIR_FREE((*group)->path); + VIR_FREE(*group); +} + + int virCgroupDelThread(virCgroupPtr cgroup, virCgroupThreadName nameval, -- 2.20.1

2 1

[libvirt] [PATCH] storage: escape ipv6 for ceph mon hosts to librados
by winhong-yili 25 Apr '19

25 Apr '19

Hosts for rbd are ceph monitor daemons. These have fixed IP addresses, so they are often referenced by IP rather than hostname for convenience, or to avoid relying on DNS. Using IPv4 addresses as the host name works already, but IPv6 addresses require rbd-specific escaping because the colon is used as an option separator in the string passed to librados. Escape these colons, and enclose the IPv6 address in square brackets so it is distinguished from the port, which is currently mandatory. Signed-off-by: Yi Li <yili(a)winhong.com> --- docs/schemas/storagepool.rng | 5 ++++- src/storage/storage_backend_rbd.c | 13 ++++++++++--- tests/storagepoolxml2xmlin/pool-rbd-ipv6.xml | 13 +++++++++++++ tests/storagepoolxml2xmlout/pool-rbd-ipv6.xml | 16 ++++++++++++++++ tests/storagepoolxml2xmltest.c | 1 + 5 files changed, 44 insertions(+), 4 deletions(-) create mode 100644 tests/storagepoolxml2xmlin/pool-rbd-ipv6.xml create mode 100644 tests/storagepoolxml2xmlout/pool-rbd-ipv6.xml diff --git a/docs/schemas/storagepool.rng b/docs/schemas/storagepool.rng index 3ca8e79..976a02b 100644 --- a/docs/schemas/storagepool.rng +++ b/docs/schemas/storagepool.rng @@ -305,7 +305,10 @@ <oneOrMore> <element name='host'> <attribute name='name'> - <text/> + <choice> + <ref name="dnsName"/> + <ref name="ipAddr"/> + </choice> </attribute> <optional> <attribute name='port'> diff --git a/src/storage/storage_backend_rbd.c b/src/storage/storage_backend_rbd.c index f8c968e..3056563 100644 --- a/src/storage/storage_backend_rbd.c +++ b/src/storage/storage_backend_rbd.c @@ -268,9 +268,16 @@ virStorageBackendRBDOpenRADOSConn(virStorageBackendRBDStatePtr ptr, source->hosts[i].name); } else if (source->hosts[i].name != NULL && source->hosts[i].port) { - virBufferAsprintf(&mon_host, "%s:%d,", - source->hosts[i].name, - source->hosts[i].port); + /* assume host containing : is ipv6 */ + if (strchr(source->hosts[i].name, ':')) { + virBufferAsprintf(&mon_host, "[%s]:%d,", + source->hosts[i].name, + source->hosts[i].port); + } else { + virBufferAsprintf(&mon_host, "%s:%d,", + source->hosts[i].name, + source->hosts[i].port); + } } else { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("received malformed monitor, check the XML definition")); diff --git a/tests/storagepoolxml2xmlin/pool-rbd-ipv6.xml b/tests/storagepoolxml2xmlin/pool-rbd-ipv6.xml new file mode 100644 index 0000000..0744b33 --- /dev/null +++ b/tests/storagepoolxml2xmlin/pool-rbd-ipv6.xml @@ -0,0 +1,13 @@ +<pool type='rbd'> + <name>ceph</name> + <uuid>47c1faee-0207-e741-f5ae-d9b019b98fe2</uuid> + <source> + <name>rbd</name> + <host name='localhost' port='6789'/> + <host name='localhost' port='6790'/> + <host name='2205::192:168:205:141' port='6789'/> + <auth username='admin' type='ceph'> + <secret uuid='2ec115d7-3a88-3ceb-bc12-0ac909a6fd87'/> + </auth> + </source> +</pool> diff --git a/tests/storagepoolxml2xmlout/pool-rbd-ipv6.xml b/tests/storagepoolxml2xmlout/pool-rbd-ipv6.xml new file mode 100644 index 0000000..cc2a379 --- /dev/null +++ b/tests/storagepoolxml2xmlout/pool-rbd-ipv6.xml @@ -0,0 +1,16 @@ +<pool type='rbd'> + <name>ceph</name> + <uuid>47c1faee-0207-e741-f5ae-d9b019b98fe2</uuid> + <capacity unit='bytes'>0</capacity> + <allocation unit='bytes'>0</allocation> + <available unit='bytes'>0</available> + <source> + <host name='localhost' port='6789'/> + <host name='localhost' port='6790'/> + <host name='2205::192:168:205:141' port='6789'/> + <name>rbd</name> + <auth type='ceph' username='admin'> + <secret uuid='2ec115d7-3a88-3ceb-bc12-0ac909a6fd87'/> + </auth> + </source> +</pool> diff --git a/tests/storagepoolxml2xmltest.c b/tests/storagepoolxml2xmltest.c index 2ae514f..b6f4cb4 100644 --- a/tests/storagepoolxml2xmltest.c +++ b/tests/storagepoolxml2xmltest.c @@ -95,6 +95,7 @@ mymain(void) DO_TEST("pool-zfs-sourcedev"); DO_TEST("pool-rbd"); #ifdef WITH_STORAGE_RBD + DO_TEST("pool-rbd-ipv6"); DO_TEST("pool-rbd-refresh-volume-allocation"); DO_TEST("pool-rbd-ns-configopts"); #endif -- 2.7.5

1 0

[libvirt] [PATCH] conf: add cpu check attribute to ABI check
by Nikolay Shirokovskiy 24 Apr '19

24 Apr '19

Different check values is not ABI compatible. For example if on migration we change 'full' to 'partial' then guest cpu on destination can be different. Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy(a)virtuozzo.com> --- src/conf/cpu_conf.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/src/conf/cpu_conf.c b/src/conf/cpu_conf.c index 33c8b99..bd2beab 100644 --- a/src/conf/cpu_conf.c +++ b/src/conf/cpu_conf.c @@ -894,6 +894,13 @@ virCPUDefIsEqual(virCPUDefPtr src, goto cleanup; } + if (src->check != dst->check) { + MISMATCH(_("Target CPU check %s does not match source %s"), + virCPUCheckTypeToString(dst->check), + virCPUCheckTypeToString(src->check)); + goto cleanup; + } + if (src->arch != dst->arch) { MISMATCH(_("Target CPU arch %s does not match source %s"), virArchToString(dst->arch), -- 1.8.3.1

2 1

[libvirt] JVM crashes during GC
by Sachin Soman 24 Apr '19

24 Apr '19

Hi, Could you tell me if the following is some known issue? While performing the following simple test, I see my JVM crashing (consistently): 1. Open a connection to an ESXi driver/host (passing ConnectAuthDefault instance). 2. Close the connection. 3. Invoke GC When GC is triggered, at some point, some unallocated native memory is being tried to release. That's failing. The error thrown is: java(78745,0x70000241e000) malloc: *** error for object 0x7fd5df561390: pointer being freed was not allocated *** set a breakpoint in malloc_error_break to debug Frames from core dump: frame #0: 0x00007fff5b274b66 libsystem_kernel.dylib`__pthread_kill + 10 frame #1: 0x00007fff5b43f080 libsystem_pthread.dylib`pthread_kill + 333 frame #2: 0x00007fff5b1d01ae libsystem_c.dylib`abort + 127 frame #3: 0x00007fff5b2ce8a6 libsystem_malloc.dylib`free + 521 frame #4: 0x00000001127f43a7 frame #5: 0x00000001127e3ffd frame #6: 0x00000001127e3ffd frame #7: 0x00000001127e3ffd frame #8: 0x00000001127e3ffd frame #9: 0x00000001127e4042 frame #10: 0x00000001127e3ffd frame #11: 0x00000001127e3ffd frame #12: 0x00000001127dc4e7 frame #13: 0x000000010c0e235e libjvm.dylib`JavaCalls::call_helper(JavaValue*, methodHandle*, JavaCallArguments*, Thread*) + 1710 frame #14: 0x000000010c0e2b02 libjvm.dylib`JavaCalls::call_virtual(JavaValue*, KlassHandle, Symbol*, Symbol*, JavaCallArguments*, Thread*) + 356 frame #15: 0x000000010c0e2cae libjvm.dylib`JavaCalls::call_virtual(JavaValue*, Handle, KlassHandle, Symbol*, Symbol*, Thread*) + 74 frame #16: 0x000000010c1208ee libjvm.dylib`thread_entry(JavaThread*, Thread*) + 124 frame #17: 0x000000010c33e84d libjvm.dylib`JavaThread::thread_main_inner() + 155 frame #18: 0x000000010c33ff12 libjvm.dylib`JavaThread::run() + 448 frame #19: 0x000000010c26058a libjvm.dylib`java_start(Thread*) + 246 frame #20: 0x00007fff5b43c661 libsystem_pthread.dylib`_pthread_body + 340 frame #21: 0x00007fff5b43c50d libsystem_pthread.dylib`_pthread_start + 377 frame #22: 0x00007fff5b43bbf9 libsystem_pthread.dylib`thread_start + 13 I have installed Libvirt 5.2.0. Java bindings libvirt-java 0.5.1 JNA 4.0.0 Tested Java environments: Oracle Java 8 and OpenJDK 8 on MAC, OpenJDK 11 on Ubuntu 16 Thanks & Regards Sachin Soman

3 11

[libvirt] [PATCH] networkStartNetworkVirtual: Don't overwrite error in 'err5'
by Michal Privoznik 23 Apr '19

23 Apr '19

If there's an error when setting up QoS on a bridge the control jumps over to 'err5' label. Here, the virNetDevBandwidthClear() is called to clear out any partially set QoS. This function can also report an error which would overwrite the actual error that caused us jumping here. 🤦 Use virErrorPreserveLast() to preserve the original error. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/network/bridge_driver.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/network/bridge_driver.c b/src/network/bridge_driver.c index ce4f4890f1..77206b4584 100644 --- a/src/network/bridge_driver.c +++ b/src/network/bridge_driver.c @@ -2493,6 +2493,7 @@ networkStartNetworkVirtual(virNetworkDriverStatePtr driver, return 0; err5: + virErrorPreserveLast(&save_err); if (def->bandwidth) virNetDevBandwidthClear(def->bridge); -- 2.21.0

2 3

[libvirt] [PATCH 0/3] Emit event on lease attach/detach
by Michal Privoznik 23 Apr '19

23 Apr '19

Unfortunately, we can't emit VIR_DOMAIN_EVENT_ID_DEVICE_ADDED or VIR_DOMAIN_EVENT_ID_DEVICE_REMOVED event because that carries device alias within itself and leases don't have one. Michal Prívozník (3): qemuDomainAttachDeviceLive: Rework event emitting Introduce VIR_DOMAIN_EVENT_ID_LEASE_CHANGE qemu: Emit event on lease attach/detach examples/object-events/event-test.c | 34 +++++++++++ include/libvirt/libvirt-domain.h | 33 +++++++++++ src/conf/domain_event.c | 89 +++++++++++++++++++++++++++++ src/conf/domain_event.h | 12 ++++ src/libvirt_private.syms | 2 + src/qemu/qemu_driver.c | 43 +++++++------- src/qemu/qemu_hotplug.c | 7 +++ src/remote/remote_daemon_dispatch.c | 40 +++++++++++++ src/remote/remote_driver.c | 32 +++++++++++ src/remote/remote_protocol.x | 16 +++++- src/remote_protocol-structs | 8 +++ tools/virsh-domain.c | 31 ++++++++++ 12 files changed, 325 insertions(+), 22 deletions(-) -- 2.21.0

3 7