October 2020 - Devel - libvirt List Archives

[PATCH v2 0/7] more Hyper-V APIs
by Matt Coleman 19 Oct '20

19 Oct '20

This set of patches adds several new APIs and fixes several others. Changes since v1: * 'enabledValue' and 'disabledValue' are now static strings in hypervDomainSetAutostart() * a long virReportError() call has been split into two lines Matt Coleman (7): hyperv: implement domainSetAutostart hyperv: implement nodeGetFreeMemory hyperv: implement domainReboot and domainReset hyperv: implement domainShutdown and domainShutdownFlags hyperv: fix domainSuspend and domainResume on Hyper-V V2 hyperv: fix domainManagedSave on Hyper-V V2 news: more Hyper-V APIs NEWS.rst | 7 +- src/hyperv/hyperv_driver.c | 267 +++++++++++++++++++++++++- src/hyperv/hyperv_wmi_classes.h | 3 + src/hyperv/hyperv_wmi_generator.input | 78 ++++++++ 4 files changed, 347 insertions(+), 8 deletions(-) -- 2.27.0

4 24

[PATCH 0/5] hyperv: Make ownershipt transfer obvious
by Michal Privoznik 19 Oct '20

19 Oct '20

*** BLURB HERE *** Michal Prívozník (5): hyperv: Make it obvious that hypervInvokeMethod() consumes an argument hyperv: Reindent hypervInvokeMethod() body hyperv: Drop needless label in hypervDomainSetMemoryFlags() hyperv: Make it obvious that hypervAddEmbeddedParam() consumes an argument hyperv: Simplify @memResource freeing in hypervDomainSetMemoryFlags() src/hyperv/hyperv_driver.c | 34 ++++------ src/hyperv/hyperv_wmi.c | 129 +++++++++++++++++++++---------------- src/hyperv/hyperv_wmi.h | 12 ++-- 3 files changed, 94 insertions(+), 81 deletions(-) -- 2.26.2

3 7

Re: [PATCH 0/6] qemu: Fix cdrom as SCSI hostdev via -blockdev
by daggs 19 Oct '20

19 Oct '20

Greetings All, > Sent: Thursday, October 15, 2020 at 6:25 PM > From: "Peter Krempa" <pkrempa(a)redhat.com> > To: libvir-list(a)redhat.com > Cc: "daggs" <daggs(a)gmx.com> > Subject: [PATCH 0/6] qemu: Fix cdrom as SCSI hostdev via -blockdev > > -drive detects whether a device is a cdrom automatically but we need to > use 'host_cdrom' when using blockdev explicitly. Fix the hostdev code > which was recently converted to -blockdev. > > Warning: > > I _don't_ have a box with a CDROM handy so this code is not tested in action! > > Peter Krempa (6): > qemuProcessCreatePretendCmd: Split up preparation and command building > qemu: hostdev: Prepare definition bits in qemuDomainPrepareHostdev > qemu: Prepare hostdev data which depends on the host state separately > qemuxml2argvmock: Remove mocking of 'virSCSIDeviceGetSgName' > qemu: Detect whether a SCSI hostdev is a cdrom > qemu: Add test cases for 'host_cdrom' blockdev backend via <disk> I've tested the bellow on libvirt 6.8.0, I wasn't able to get it to work, with the same entry as the example, e.g.: <hostdev mode='subsystem' type='scsi' managed='yes'> <source> <adapter name='cdrom'/> <address bus='0' target='0' unit='0'/> </source> <readonly/> <address type='drive' controller='0' bus='0' target='0' unit='0'/> </hostdev> I get this error: error: XML error: SCSI host device doesn't support managed mode. if I mark the managed as no (like other pci pass-through devices I have in the vm), I get this error: error: internal error: Cannot parse adapter 'cdrom' looks like the commit is missing the cdrom validation at https://gitlab.com/libvirt/libvirt/-/blob/master/src/util/virscsi.c#L90 full machine xml can be found at https://dpaste.com/G32WWAJA2 Dagg.

1 0

[libvirt PATCH v2 0/4] Make SEV 'cbitpos' and 'reducedPhysBits' attributes optional
by Erik Skultety 16 Oct '20

16 Oct '20

We designed them as mandatory, but these are platform dependent and can be filled from QEMU capabilities. Resolves: https://gitlab.com/libvirt/libvirt/-/issues/57 Since v1: - moved the capability check from qemuProcessPrepareSEVGuestInput to qemuValidateDomainDef as requested Erik Skultety (4): qemu_process: sev: Drop an unused variable qemu: process: Move SEV capability check to qemuValidateDomainDef qemu: process: sev: Fill missing 'cbitpos' & 'reducedPhysBits' from caps conf: domain: sev: Make 'cbitpos' & 'reducedPhysBits' attrs optional docs/schemas/domaincommon.rng | 16 ++++--- src/conf/domain_conf.c | 46 ++++++++++++------- src/conf/domain_conf.h | 2 + src/qemu/qemu_process.c | 45 +++++++++++++----- src/qemu/qemu_validate.c | 8 ++++ ...v-missing-platform-info.x86_64-2.12.0.args | 37 +++++++++++++++ ...nch-security-sev-missing-platform-info.xml | 35 ++++++++++++++ tests/qemuxml2argvtest.c | 1 + 8 files changed, 157 insertions(+), 33 deletions(-) create mode 100644 tests/qemuxml2argvdata/launch-security-sev-missing-platfo= rm-info.x86_64-2.12.0.args create mode 100644 tests/qemuxml2argvdata/launch-security-sev-missing-platfo= rm-info.xml --=20 2.26.2

2 6

[RFC PATCH v2] fix error message in virMigrate3 if connection is broken
by Nikolay Shirokovskiy 15 Oct '20

15 Oct '20

Changes from v1 [1]: - return error value from VIR_DRV_SUPPORTS_FEATURE instead of setting error to out argument (decrease over enginering level on patch generator in other words :) Currently virDomainMigrate3 reports "this function is not supported by the connection driver: virDomainMigrate3" if connection to destination for example is broken. This is a bit misleading. This is a result of we treat errors as unsupported feature in VIR_DRV_SUPPORTS_FEATURE macro. Let's return error from macro as is. I guess all the other occurences of VIR_DRV_SUPPORTS_FEATURE need to be updated as well so that we detect errors early and not have issues similar to virDomainMigrate3. I'm going to fix the other places if this RFC is approved. [1] https://www.redhat.com/archives/libvir-list/2020-September/msg01056.html --- src/driver.h | 14 ++++++++ src/libvirt-domain.c | 91 +++++++++++++++++++++++++++++++++++++--------------- 2 files changed, 79 insertions(+), 26 deletions(-) diff --git a/src/driver.h b/src/driver.h index 6278aa0..c29b2fa 100644 --- a/src/driver.h +++ b/src/driver.h @@ -60,6 +60,20 @@ typedef enum { (drv)->connectSupportsFeature((conn), (feature)) > 0 : 0) +/* + * A little wrapper for connectSupportsFeature API to test that the API + * itself is available first. Return values are same as for API. + * + * Returns: + * -1 on error + * 0 feature is not supported + * 1 feature is supported + */ +#define VIR_DRV_SUPPORTS_FEATURE2(drv, conn, feature) \ + ((drv)->connectSupportsFeature ? \ + (drv)->connectSupportsFeature((conn), (feature)) : 0) + + #define __VIR_DRIVER_H_INCLUDES___ #include "driver-hypervisor.h" diff --git a/src/libvirt-domain.c b/src/libvirt-domain.c index cde86c7..03c357f 100644 --- a/src/libvirt-domain.c +++ b/src/libvirt-domain.c @@ -3846,6 +3846,9 @@ virDomainMigrate3(virDomainPtr domain, const char *dname = NULL; const char *dxml = NULL; unsigned long long bandwidth = 0; + int rc_src; + int rc_dst; + int rc; VIR_DOMAIN_DEBUG(domain, "dconn=%p, params=%p, nparms=%u flags=0x%x", dconn, params, nparams, flags); @@ -3878,15 +3881,21 @@ virDomainMigrate3(virDomainPtr domain, } if (flags & VIR_MIGRATE_OFFLINE) { - if (!VIR_DRV_SUPPORTS_FEATURE(domain->conn->driver, domain->conn, - VIR_DRV_FEATURE_MIGRATION_OFFLINE)) { + rc = VIR_DRV_SUPPORTS_FEATURE2(domain->conn->driver, domain->conn, + VIR_DRV_FEATURE_MIGRATION_OFFLINE); + if (rc < 0) { + goto error; + } else if (rc == 0) { virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s", _("offline migration is not supported by " "the source host")); goto error; } - if (!VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn, - VIR_DRV_FEATURE_MIGRATION_OFFLINE)) { + rc = VIR_DRV_SUPPORTS_FEATURE2(dconn->driver, dconn, + VIR_DRV_FEATURE_MIGRATION_OFFLINE); + if (rc < 0) { + goto error; + } else if (rc == 0) { virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s", _("offline migration is not supported by " "the destination host")); @@ -3899,21 +3908,30 @@ virDomainMigrate3(virDomainPtr domain, * the flag for just the source side. We mask it out to allow * migration from newer source to an older destination that * rejects the flag. */ - if (flags & VIR_MIGRATE_CHANGE_PROTECTION && - !VIR_DRV_SUPPORTS_FEATURE(domain->conn->driver, domain->conn, - VIR_DRV_FEATURE_MIGRATE_CHANGE_PROTECTION)) { - virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s", - _("cannot enforce change protection")); - goto error; + if (flags & VIR_MIGRATE_CHANGE_PROTECTION) { + rc = VIR_DRV_SUPPORTS_FEATURE2(domain->conn->driver, domain->conn, + VIR_DRV_FEATURE_MIGRATE_CHANGE_PROTECTION); + if (rc < 0) { + goto error; + } else if (rc == 0) { + virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s", + _("cannot enforce change protection")); + goto error; + } } flags &= ~VIR_MIGRATE_CHANGE_PROTECTION; /* Prefer extensible API but fall back to older migration APIs if params * only contains parameters which were supported by the older API. */ - if (VIR_DRV_SUPPORTS_FEATURE(domain->conn->driver, domain->conn, - VIR_DRV_FEATURE_MIGRATION_PARAMS) && - VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn, - VIR_DRV_FEATURE_MIGRATION_PARAMS)) { + rc_src = VIR_DRV_SUPPORTS_FEATURE2(domain->conn->driver, domain->conn, + VIR_DRV_FEATURE_MIGRATION_PARAMS); + if (rc_src < 0) + goto error; + rc_dst = VIR_DRV_SUPPORTS_FEATURE2(dconn->driver, dconn, + VIR_DRV_FEATURE_MIGRATION_PARAMS); + if (rc_dst < 0) + goto error; + if (rc_src && rc_dst) { VIR_DEBUG("Using migration protocol 3 with extensible parameters"); ddomain = virDomainMigrateVersion3Params(domain, dconn, params, nparams, flags); @@ -3939,17 +3957,30 @@ virDomainMigrate3(virDomainPtr domain, goto error; } - if (VIR_DRV_SUPPORTS_FEATURE(domain->conn->driver, domain->conn, - VIR_DRV_FEATURE_MIGRATION_V3) && - VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn, - VIR_DRV_FEATURE_MIGRATION_V3)) { + rc_src = VIR_DRV_SUPPORTS_FEATURE2(domain->conn->driver, domain->conn, + VIR_DRV_FEATURE_MIGRATION_V3); + if (rc_src < 0) + goto error; + rc_dst = VIR_DRV_SUPPORTS_FEATURE2(dconn->driver, dconn, + VIR_DRV_FEATURE_MIGRATION_V3); + if (rc_dst < 0) + goto error; + if (rc_src && rc_dst) { VIR_DEBUG("Using migration protocol 3"); ddomain = virDomainMigrateVersion3(domain, dconn, dxml, flags, dname, uri, bandwidth); - } else if (VIR_DRV_SUPPORTS_FEATURE(domain->conn->driver, domain->conn, - VIR_DRV_FEATURE_MIGRATION_V2) && - VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn, - VIR_DRV_FEATURE_MIGRATION_V2)) { + goto done; + } + + rc_src = VIR_DRV_SUPPORTS_FEATURE2(domain->conn->driver, domain->conn, + VIR_DRV_FEATURE_MIGRATION_V2); + if (rc_src < 0) + goto error; + rc_dst = VIR_DRV_SUPPORTS_FEATURE2(dconn->driver, dconn, + VIR_DRV_FEATURE_MIGRATION_V2); + if (rc_dst < 0) + goto error; + if (rc_src && rc_dst) { VIR_DEBUG("Using migration protocol 2"); if (dxml) { virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s", @@ -3959,10 +3990,18 @@ virDomainMigrate3(virDomainPtr domain, } ddomain = virDomainMigrateVersion2(domain, dconn, flags, dname, uri, bandwidth); - } else if (VIR_DRV_SUPPORTS_FEATURE(domain->conn->driver, domain->conn, - VIR_DRV_FEATURE_MIGRATION_V1) && - VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn, - VIR_DRV_FEATURE_MIGRATION_V1)) { + goto done; + } + + rc_src = VIR_DRV_SUPPORTS_FEATURE2(domain->conn->driver, domain->conn, + VIR_DRV_FEATURE_MIGRATION_V1); + if (rc_src < 0) + goto error; + rc_dst = VIR_DRV_SUPPORTS_FEATURE2(dconn->driver, dconn, + VIR_DRV_FEATURE_MIGRATION_V1); + if (rc_dst < 0) + goto error; + if (rc_src && rc_dst) { VIR_DEBUG("Using migration protocol 1"); if (dxml) { virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s", -- 1.8.3.1

1 1

[libvirt PATCH 0/3] Add script to sync from QEMU i386 cpu models
by Tim Wiederhake 15 Oct '20

15 Oct '20

This hopefully makes synchronization with QEMU faster and less error prone. Patch #3 showcases the changes to the cpu models the script suggests for qemu 8d90bfc5c31ad60f6049dd39be636b06bc00b652. Tim Wiederhake (3): cpu_map: Unify apostrophe and quotation mark usage cpu_map: Add script to sync from QEMU i386 cpu models [DONTMERGE] Sample output of new sync script src/cpu_map/arm_Falkor.xml | 6 +- src/cpu_map/arm_Kunpeng-920.xml | 6 +- src/cpu_map/arm_ThunderX299xx.xml | 6 +- src/cpu_map/arm_cortex-a53.xml | 6 +- src/cpu_map/arm_cortex-a57.xml | 6 +- src/cpu_map/arm_cortex-a72.xml | 6 +- src/cpu_map/arm_features.xml | 34 +- src/cpu_map/index.xml | 12 +- src/cpu_map/ppc64_POWER6.xml | 6 +- src/cpu_map/ppc64_POWER7.xml | 8 +- src/cpu_map/ppc64_POWER8.xml | 10 +- src/cpu_map/ppc64_POWER9.xml | 6 +- src/cpu_map/ppc64_POWERPC_e5500.xml | 6 +- src/cpu_map/ppc64_POWERPC_e6500.xml | 6 +- src/cpu_map/ppc64_vendors.xml | 4 +- src/cpu_map/sync_qemu_i386.py | 361 +++++++++ src/cpu_map/x86_486.xml | 18 +- src/cpu_map/x86_Broadwell-IBRS.xml | 135 ++-- src/cpu_map/x86_Broadwell-noTSX-IBRS.xml | 131 ++-- src/cpu_map/x86_Broadwell-noTSX.xml | 129 ++-- src/cpu_map/x86_Broadwell.xml | 132 ++-- src/cpu_map/x86_Cascadelake-Server-noTSX.xml | 166 ++-- src/cpu_map/x86_Cascadelake-Server.xml | 169 ++-- src/cpu_map/x86_Conroe.xml | 70 +- src/cpu_map/x86_Cooperlake.xml | 14 +- src/cpu_map/x86_Denverton.xml | 74 ++ src/cpu_map/x86_Dhyana.xml | 144 ++-- src/cpu_map/x86_EPYC-IBPB.xml | 157 ++-- src/cpu_map/x86_EPYC-Rome.xml | 169 ++-- src/cpu_map/x86_EPYC.xml | 150 ++-- src/cpu_map/x86_Haswell-IBRS.xml | 128 ++-- src/cpu_map/x86_Haswell-noTSX-IBRS.xml | 124 +-- src/cpu_map/x86_Haswell-noTSX.xml | 122 +-- src/cpu_map/x86_Haswell.xml | 124 +-- src/cpu_map/x86_Icelake-Client-noTSX.xml | 164 ++-- src/cpu_map/x86_Icelake-Client.xml | 167 ++-- src/cpu_map/x86_Icelake-Server-noTSX.xml | 197 ++--- src/cpu_map/x86_Icelake-Server.xml | 185 ++--- src/cpu_map/x86_IvyBridge-IBRS.xml | 109 +-- src/cpu_map/x86_IvyBridge.xml | 106 +-- src/cpu_map/x86_KnightsMill.xml | 77 ++ src/cpu_map/x86_Nehalem-IBRS.xml | 84 +- src/cpu_map/x86_Nehalem.xml | 81 +- src/cpu_map/x86_Opteron_G1.xml | 65 +- src/cpu_map/x86_Opteron_G2.xml | 72 +- src/cpu_map/x86_Opteron_G3.xml | 82 +- src/cpu_map/x86_Opteron_G4.xml | 105 +-- src/cpu_map/x86_Opteron_G5.xml | 111 +-- src/cpu_map/x86_Penryn.xml | 74 +- src/cpu_map/x86_SandyBridge-IBRS.xml | 98 +-- src/cpu_map/x86_SandyBridge.xml | 95 +-- src/cpu_map/x86_Skylake-Client-IBRS.xml | 148 ++-- src/cpu_map/x86_Skylake-Client-noTSX-IBRS.xml | 144 ++-- src/cpu_map/x86_Skylake-Client.xml | 145 ++-- src/cpu_map/x86_Skylake-Server-IBRS.xml | 158 ++-- src/cpu_map/x86_Skylake-Server-noTSX-IBRS.xml | 155 ++-- src/cpu_map/x86_Skylake-Server.xml | 156 ++-- src/cpu_map/x86_Snowridge.xml | 79 ++ src/cpu_map/x86_Westmere-IBRS.xml | 85 +- src/cpu_map/x86_Westmere.xml | 84 +- src/cpu_map/x86_athlon.xml | 60 +- src/cpu_map/x86_core2duo.xml | 72 +- src/cpu_map/x86_coreduo.xml | 62 +- src/cpu_map/x86_cpu64-rhel5.xml | 54 +- src/cpu_map/x86_cpu64-rhel6.xml | 58 +- src/cpu_map/x86_features.xml | 724 +++++++++--------- src/cpu_map/x86_kvm32.xml | 57 +- src/cpu_map/x86_kvm64.xml | 65 +- src/cpu_map/x86_n270.xml | 66 +- src/cpu_map/x86_pentium.xml | 31 +- src/cpu_map/x86_pentium2.xml | 49 +- src/cpu_map/x86_pentium3.xml | 51 +- src/cpu_map/x86_pentiumpro.xml | 38 +- src/cpu_map/x86_phenom.xml | 83 +- src/cpu_map/x86_qemu32.xml | 48 +- src/cpu_map/x86_qemu64.xml | 75 +- src/cpu_map/x86_vendors.xml | 6 +- 77 files changed, 4160 insertions(+), 3110 deletions(-) create mode 100755 src/cpu_map/sync_qemu_i386.py create mode 100644 src/cpu_map/x86_Denverton.xml create mode 100644 src/cpu_map/x86_KnightsMill.xml create mode 100644 src/cpu_map/x86_Snowridge.xml -- 2.26.2

2 4

[PATCH v2] docs: Document camelCase preference for XML elements and attributes
by Michal Privoznik 15 Oct '20

15 Oct '20

Recently I've merged a patch that used hyphens in an attribute name. I fixed it later, but turned out we don't document our preference which is camelCase. Suggested-by: Erik Skultety <eskultet(a)redhat.com> Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- diff to v1: - Extended coding style too. docs/api_extension.html.in | 6 ++++++ docs/coding-style.rst | 15 +++++++++++++++ 2 files changed, 21 insertions(+) diff --git a/docs/api_extension.html.in b/docs/api_extension.html.in index 6c64e83314..d7696056ac 100644 --- a/docs/api_extension.html.in +++ b/docs/api_extension.html.in @@ -110,6 +110,12 @@ src/libvirt_public.syms </code></p> + <p> + Please note that single word names for attributes and elements is + preferred. But if you have to use two ore more words please join them in + camelCase style. + </p> + <p> This task is in many ways the most important to get right, since once the API has been committed to the repository, it's libvirt's policy diff --git a/docs/coding-style.rst b/docs/coding-style.rst index 44e5265a60..cfd7b16638 100644 --- a/docs/coding-style.rst +++ b/docs/coding-style.rst @@ -960,3 +960,18 @@ git): cleanup: /* ... do other stuff ... */ } + + +XML element and attribute naming +-------------------------------- + +New elements and/or attributes should be short and descriptive. +In general, they should reflect what the feature does instead of +how exactly it is named in given hypervisor because this creates +an abstraction that other drivers can benefit from (for instance +if the same feature is named differently in two hypervisors). +That is not to say an element or attribute can't have the same +name as in a hypervisor, but proceed with caution. + +Single worded names are preferred, but if more words must be +used then they shall be joined in camelCase style. -- 2.26.2

2 1

[PATCH v1 00/24] move more validations to qemu_validate.c
by Daniel Henrique Barboza 15 Oct '20

15 Oct '20

Hi, This is another step following up the work done in [1]. Recapping, the idea is to move any validation that can be done in parse/define time to qemu_validate.c instead of let it sit inside a command line build function in qemu_command.c. I'll again copy/paste the reasoning Cole Robinson gave for this work back then (see [2] for more info): ------- The benefits of moving to validate time is that XML is rejected by 'virsh define' rather than at 'virsh start' time. It also makes it easier to follow the cli building code, and makes it easier to verify qemu_command.c test suite code coverage for the important stuff like covering every CLI option. It's also a good intermediate step for sharing validation with domain capabilities building, like is done presently for rng models. ------- Not all validations were moved with these series, but I covered most of them. One thing worth noticing is the existence of 'post start' validations, most of them related to PCI addressing, and other notable cases such as NUMA nodes and CPU models, that can't be moved to parse time. The reason is that the QEMU driver will change them during start time. I contemplated move them to the existing qemuBuildCommandLineValidate(), but that would remove them from the hotplug path that uses the qemuBuild*DevStr() functions. I don't have an easy answer on how to handle these 'post start' validations ATM, so for now they remain in qemu_command.c. [1] https://www.redhat.com/archives/libvir-list/2019-December/msg00570.html [2] https://www.redhat.com/archives/libvir-list/2019-October/msg00568.html src/qemu/qemu_command.c | 379 +-------- src/qemu/qemu_command.h | 4 +- src/qemu/qemu_hotplug.c | 4 +- src/qemu/qemu_validate.c | 740 ++++++++++++++---- tests/qemuhotplugtest.c | 1 + .../disk-sata-incompatible-address.err | 2 +- .../net-virtio-rxqueuesize-invalid-size.err | 2 +- .../pseries-panic-address.err | 2 +- tests/qemuxml2argvtest.c | 28 +- tests/qemuxml2xmltest.c | 62 +- 10 files changed, 706 insertions(+), 518 deletions(-) -- 2.26.2

2 31

[libvirt PATCH] qemu: agent: fix array access
by Ján Tomko 15 Oct '20

15 Oct '20

My code movement changed the type of ifaces_ret from virDomainInterfacePtr * to virDomainInterfacePtr **, but failed to adjust the condition or dereference the array correctly. Signed-off-by: Ján Tomko <jtomko(a)redhat.com> Fixes: 6ddb1f803ea38d8d709b984fa9539e34318a9dc0 --- src/qemu/qemu_agent.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_agent.c b/src/qemu/qemu_agent.c index c9c4b034d3..4c61c5cdc6 100644 --- a/src/qemu/qemu_agent.c +++ b/src/qemu/qemu_agent.c @@ -2209,6 +2209,7 @@ qemuAgentGetAllInterfaceAddresses(virDomainInterfacePtr **ifaces_ret, size_t ifaces_count = 0; size_t i; + *ifaces_ret = NULL; /* Hash table to handle the interface alias */ ifaces_store = virHashNew(NULL); @@ -2223,9 +2224,9 @@ qemuAgentGetAllInterfaceAddresses(virDomainInterfacePtr **ifaces_ret, return ifaces_count; error: - if (ifaces_ret) { + if (*ifaces_ret) { for (i = 0; i < ifaces_count; i++) - virDomainInterfaceFree(*ifaces_ret[i]); + virDomainInterfaceFree((*ifaces_ret)[i]); } VIR_FREE(*ifaces_ret); return -1; -- 2.26.2

2 1

[PATCH v1 1/8] migration/dirtyrate: Introduce virDomainDirtyRateInfo structure
by Hao Wang 15 Oct '20

15 Oct '20

Introduce virDomainDirtyRateInfo structure used for domain's memory dirty rate query. Signed-off-by: Hao Wang <wanghao232(a)huawei.com> Reviewed-by: Chuan Zheng <zhengchuan(a)huawei.com> --- include/libvirt/libvirt-domain.h | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) diff --git a/include/libvirt/libvirt-domain.h b/include/libvirt/libvirt-domain.h index 8b9d9c1..2d3847b 100644 --- a/include/libvirt/libvirt-domain.h +++ b/include/libvirt/libvirt-domain.h @@ -4999,4 +4999,28 @@ int virDomainBackupBegin(virDomainPtr domain, char *virDomainBackupGetXMLDesc(virDomainPtr domain, unsigned int flags); +/** + * virDomainDirtyRateInfo: + * + * a virDomainDirtyRateInfo is a structure filled by virDomainGetDirtyRate() and + * extracting dirty rate infomation for a given active Domain. + */ + +typedef struct _virDomainDirtyRateInfo virDomainDirtyRateInfo; + +struct _virDomainDirtyRateInfo { + int status; /* the status of dirtyrate calculation */ + long long dirtyRate; /* the dirtyrate in MB/s */ + long long startTime; /* the start time of dirtyrate calculation */ + long long calcTime; /* the period of dirtyrate calculation */ +}; + +/** + * virDomainDirtyRateInfoPtr: + * + * a virDomainDirtyRateInfoPtr is a pointer to a virDomainDirtyRateInfo structure. + */ + +typedef virDomainDirtyRateInfo *virDomainDirtyRateInfoPtr; + #endif /* LIBVIRT_DOMAIN_H */ -- 2.23.0

3 9