- Devel - libvirt List Archives

[libvirt PATCHv2 0/5] qemu: do not use deprecated options for new virtiofsd (virtiofs epopee)
by Ján Tomko 09 Jul '24

09 Jul '24

Also, allow migration if we think we can. Ján Tomko (5): tests: vhostuser: add virtiofsd json descriptor tests: qemuxmlconf: adjust test case to new virtiofsd qemu: fill capabilities for virtiofsd qemu: do not use deprecated options for new virtiofsd qemu: migration: allow migration for virtiofs src/conf/domain_conf.h | 1 + src/qemu/qemu_migration.c | 9 +- src/qemu/qemu_vhost_user.c | 60 ++++++++++++ src/qemu/qemu_vhost_user.h | 12 +++ src/qemu/qemu_virtiofs.c | 92 +++++++++++++------ .../libexec/qemu/vhost-user/test-virtiofsd | 10 ++ .../qemu/vhost-user/50-qemu-virtiofsd.json | 5 + tests/qemuvhostusertest.c | 1 + .../vhost-user-fs-fd-memory.xml | 3 +- 9 files changed, 161 insertions(+), 32 deletions(-) create mode 100755 tests/qemuvhostuserdata/usr/libexec/qemu/vhost-user/test-virtiofsd create mode 100644 tests/qemuvhostuserdata/usr/share/qemu/vhost-user/50-qemu-virtiofsd.json -- 2.45.2

2 7

[PATCH v3 0/2] security_manager: Fix security manager deadlock after libvirtd fork()
by hongmianquan 09 Jul '24

09 Jul '24

We have the stack security driver, which internally manages other security drivers, just call them "top" and "nested". We call virSecurityStackPreFork() to lock the top one, and it also locks and then unlocks the nested drivers prior to fork. Then in qemuSecurityPostFork(), it unlocks the top one, but not the nested ones. Thus, if one of the nested drivers ("dac" or "selinux") is still locked, it will cause a deadlock. If we always surround nested locks with top lock, it is always secure. Because we have got top lock before fork child libvirtd. However, it is not always the case in the current code, We discovered this case: the nested list obtained through the qemuSecurityGetNested() will be locked directly for subsequent use, such as in virQEMUDriverCreateCapabilities(), where the nested list is locked using qemuSecurityGetDOI, but the top one is not locked beforehand. The problem stack is as follows: libvirtd thread1 libvirtd thread2 child libvirtd | | | | | | virsh capabilities qemuProcessLanuch | | | | | lock top | | | | lock nested | | | | | | fork------------------->|(nested lock held by thread1) | | | | | | unlock nested unlock top unlock top | | qemuSecuritySetSocketLabel | | lock nested (deadlock) v3 changes: Made modifications based on Michal's comments - ensured matching qemuSecurityStackLock() and qemuSecurityStackUnlock() - modify the correct order in libvirt_private.syms - split the code streamlining part into a separate patch hongmianquan (2): security_manager: Ensure top lock is acquired before nested locks security_manager: Remove redundant qemuSecurityGetNested() call src/libvirt_private.syms | 2 ++ src/qemu/qemu_conf.c | 13 +++++++++++-- src/qemu/qemu_driver.c | 21 +++++++++++++-------- src/qemu/qemu_security.h | 2 ++ src/security/security_manager.c | 22 ++++++++++++++++++++++ src/security/security_manager.h | 2 ++ 6 files changed, 52 insertions(+), 10 deletions(-) -- 2.20.1

3 5

[PATCH 0/1] qemuDomainChangeNet: check virtio options for non-virtio models
by Miroslav Los 09 Jul '24

09 Jul '24

Hi, in our project, we create all of our domain interfaces regardless of model with a <driver><host csum='off'/></driver>. The documentation is not clear as to which models support offloading or not, and the domain validates. In libvirtd, since <driver> is present, this causes each network device to have its virDomainVirtioOptions *virtio to be non-NULL, with all members ABSENT, which is also deemed valid. However, clients receive the interface elements without the driver. When the modified value is sent back via updateDeviceFlags, the candidate's virtio attribute is thus NULL. The validation in qemuDomainChangeNet requires both structs to be equal or both NULL, which is violated by such a request. This can be seen also with how virsh domif-setlink operates: virsh # domiflist 5 Interface Type Source Model MAC -------------------------------------------------------- - udp - e1000 52:54:00:1c:10:42 - udp - e1000 52:54:00:09:29:9c - udp - e1000 52:54:00:1d:a6:8d - udp - e1000 52:54:00:07:4a:83 52:54:00:1c:10:42 up virsh # domif-setlink 5 52:54:00:1c:10:42 down error: Failed to update interface link state error: Operation not supported: cannot modify virtio network device driver options A workaround is to add an empty <driver> in the update request, which works until libvirt is restarted. The driver element on a non-virtio interface is not present when the domain is reloaded, thus the pointer is NULL in olddev, causing the inverse issue and same error. Our current workaround is to retry with the empty driver element only if the first call fails, but we think libvirt should handle this struct more consistently. This patch expands the check to the non-virtio case, where the olddev is assumed to be valid, and newdev's virtio options are validated to be NULL or all-ABSENT. Since the driver element is normally preserved for virtio models, the stricter equality condition is left unchanged in that case. I did not investigate further if the discrepancy between NULL and all-ABSENT values for virDomainVirtioOptions can cause a problem elsewhere, or in other device kinds. Though I do wonder if the options handling should be improved. One possibility is that non-virtio devices keep the value NULL, e.g. after validating that the driver element does not contain any of the options, whereas virtio devices always create the struct even with no driver element present on input. Another option is that NULL pointer is treated the same as all-ABSENT struct, e.g. in virDomainCheckVirtioOptionsAreAbsent and virDomainVirtioOptionsCheckABIStability; the former does that while the latter does not. These functions would also need to become callable from the qemu driver. Miroslav Los (1): qemuDomainChangeNet: check virtio options for non-virtio models src/qemu/qemu_hotplug.c | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-) -- 2.25.1

2 2

[PATCH] virt-aa-helper: Allow RO access to /usr/share/edk2-ovmf
by Michal Privoznik 09 Jul '24

09 Jul '24

When binary version of edk2 is distributed, the files reside under /usr/share/edk2-ovmf as can be seen from Gentoo's ebuild [1]. Allow virt-aa-helper to generate paths under that dir. 1: https://gitweb.gentoo.org/repo/gentoo.git/tree/sys-firmware/edk2-ovmf-bin/e… Resolves: https://bugs.gentoo.org/911786 Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/security/virt-aa-helper.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/security/virt-aa-helper.c b/src/security/virt-aa-helper.c index 402cbd9602..076b98a1d7 100644 --- a/src/security/virt-aa-helper.c +++ b/src/security/virt-aa-helper.c @@ -475,6 +475,7 @@ valid_path(const char *path, const bool readonly) "/initrd", "/initrd.img", "/usr/share/edk2/", + "/usr/share/edk2-ovmf/", "/usr/share/OVMF/", /* for OVMF images */ "/usr/share/ovmf/", /* for OVMF images */ "/usr/share/AAVMF/", /* for AAVMF images */ -- 2.44.2

3 2

[PATCH 0/6] qemu: Fixes to firmware selection
by Andrea Bolognani 09 Jul '24

09 Jul '24

Andrea Bolognani (6): tests: Update firmware descriptors tests: Add more firmware selection coverage qemu: Filter firmware images by type qemu: ROM firmware images are always readonly tests: Add firmware descriptor for edk2 on riscv64 tests: Add test for UEFI autoselection on riscv64 src/qemu/qemu_firmware.c | 17 +++++++++ .../qemu_5.2.0-tcg-virt.riscv64.xml | 4 ++- .../qemu_5.2.0-virt.riscv64.xml | 4 ++- .../qemu_8.0.0-tcg-virt.riscv64.xml | 4 ++- .../qemu_8.0.0-virt.riscv64.xml | 4 ++- .../firmware/60-edk2-ovmf-x64-inteltdx.json | 10 +++--- .../qemu/firmware/50-edk2-riscv-qcow2.json | 33 +++++++++++++++++ .../firmware/60-edk2-ovmf-x64-inteltdx.json | 10 +++--- tests/qemufirmwaretest.c | 7 +++- ...efi-riscv64.riscv64-latest.abi-update.args | 34 ++++++++++++++++++ ...-efi-riscv64.riscv64-latest.abi-update.xml | 28 +++++++++++++++ .../firmware-auto-efi-riscv64.xml | 14 ++++++++ ...ware-auto-efi-rw-pflash.x86_64-latest.err} | 0 ...mware-auto-efi-rw-pflash.x86_64-latest.xml | 35 +++++++++++++++++++ .../firmware-auto-efi-rw-pflash.xml | 18 ++++++++++ tests/qemuxmlconftest.c | 3 +- 16 files changed, 207 insertions(+), 18 deletions(-) create mode 100644 tests/qemufirmwaredata/usr/share/qemu/firmware/50-edk2-riscv-qcow2.json create mode 100644 tests/qemuxmlconfdata/firmware-auto-efi-riscv64.riscv64-latest.abi-update.args create mode 100644 tests/qemuxmlconfdata/firmware-auto-efi-riscv64.riscv64-latest.abi-update.xml create mode 100644 tests/qemuxmlconfdata/firmware-auto-efi-riscv64.xml rename tests/qemuxmlconfdata/{firmware-auto-efi-rw.x86_64-latest.abi-update.err => firmware-auto-efi-rw-pflash.x86_64-latest.err} (100%) create mode 100644 tests/qemuxmlconfdata/firmware-auto-efi-rw-pflash.x86_64-latest.xml create mode 100644 tests/qemuxmlconfdata/firmware-auto-efi-rw-pflash.xml -- 2.45.2

2 7

[PATCH] vmx: Do not require all ID data for VMWare Distributed Switch
by Martin Kletzander 08 Jul '24

08 Jul '24

Similarly to commit 2482801608b8 we can safely ignore connectionId, portId and portgroupId in both XML and VMX as they are only a blind pass-through between XML and VMX and an ethernet without such parameters was spotted in the wild. On top of that even our documentation says the whole VMWare Distrubuted Switch configuration is a best-effort. Resolves: https://issues.redhat.com/browse/RHEL-46099 Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- src/conf/domain_conf.c | 11 ++++----- src/conf/schemas/domaincommon.rng | 24 ++++++++++++------- src/vmx/vmx.c | 24 ++++++++++++------- ...-portid.vmx => ethernet-vds-no-params.vmx} | 2 -- ...-portid.xml => ethernet-vds-no-params.xml} | 2 +- 5 files changed, 37 insertions(+), 26 deletions(-) rename tests/vmx2xmldata/{ethernet-vds-no-portid.vmx => ethernet-vds-no-params.vmx} (76%) rename tests/vmx2xmldata/{ethernet-vds-no-portid.xml => ethernet-vds-no-params.xml} (82%) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 6080f4f90a54..bfef89e1beae 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -9593,15 +9593,14 @@ virDomainNetDefParseXML(virDomainXMLOption *xmlopt, def->data.vds.switch_id) < 0) return NULL; - if (virXMLPropLongLong(source_node, "portid", 0, VIR_XML_PROP_REQUIRED, - &def->data.vds.port_id, def->data.vds.port_id) < 0) + if (virXMLPropLongLong(source_node, "portid", 0, VIR_XML_PROP_NONE, + &def->data.vds.port_id, 0) < 0) return NULL; - if (!(def->data.vds.portgroup_id = virXMLPropStringRequired(source_node, "portgroupid"))) - return NULL; + def->data.vds.portgroup_id = virXMLPropString(source_node, "portgroupid"); - if (virXMLPropLongLong(source_node, "connectionid", 0, VIR_XML_PROP_REQUIRED, - &def->data.vds.connection_id, def->data.vds.connection_id) < 0) + if (virXMLPropLongLong(source_node, "connectionid", 0, VIR_XML_PROP_NONE, + &def->data.vds.connection_id, 0) < 0) return NULL; break; diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index b163e4eece09..2d23fcf12375 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -3684,15 +3684,21 @@ <attribute name="switchid"> <ref name="UUID"/> </attribute> - <attribute name="portid"> - <data type="long"/> - </attribute> - <attribute name="portgroupid"> - <data type="string"/> - </attribute> - <attribute name="connectionid"> - <data type="long"/> - </attribute> + <optional> + <attribute name="portid"> + <data type="long"/> + </attribute> + </optional> + <optional> + <attribute name="portgroupid"> + <data type="string"/> + </attribute> + </optional> + <optional> + <attribute name="connectionid"> + <data type="long"/> + </attribute> + </optional> </element> <ref name="interface-options"/> </interleave> diff --git a/src/vmx/vmx.c b/src/vmx/vmx.c index d082a0766010..e5bc2d793c66 100644 --- a/src/vmx/vmx.c +++ b/src/vmx/vmx.c @@ -2896,7 +2896,7 @@ virVMXParseEthernet(virConf *conf, int controller, virDomainNetDef **def) if (virVMXGetConfigString(conf, portgroupId_name, &(*def)->data.vds.portgroup_id, - false) < 0 || + true) < 0 || virVMXGetConfigLong(conf, portId_name, &(*def)->data.vds.port_id, @@ -2906,7 +2906,7 @@ virVMXParseEthernet(virConf *conf, int controller, virDomainNetDef **def) connectionId_name, &(*def)->data.vds.connection_id, 0, - false) < 0) + true) < 0) goto cleanup; } else if (connectionType == NULL && networkName == NULL) { (*def)->type = VIR_DOMAIN_NET_TYPE_NULL; @@ -4038,14 +4038,22 @@ virVMXFormatEthernet(virDomainNetDef *def, int controller, uuid[5], uuid[6], uuid[7], uuid[8], uuid[9], uuid[10], uuid[11], uuid[12], uuid[13], uuid[14], uuid[15]); - virBufferAsprintf(buffer, "ethernet%d.dvs.portId = \"%lld\"\n", - controller, def->data.vds.port_id); + if (def->data.vds.port_id) { + virBufferAsprintf(buffer, "ethernet%d.dvs.portId = \"%lld\"\n", + controller, def->data.vds.port_id); + } + + if (def->data.vds.portgroup_id) { + virBufferAsprintf(buffer, "ethernet%d.dvs.", controller); + virBufferEscapeString(buffer, "portgroupId = \"%s\"\n", + def->data.vds.portgroup_id); + } - virBufferAsprintf(buffer, "ethernet%d.dvs.", controller); - virBufferEscapeString(buffer, "portgroupId = \"%s\"\n", def->data.vds.portgroup_id); + if (def->data.vds.connection_id) { + virBufferAsprintf(buffer, "ethernet%d.dvs.connectionId = \"%lld\"\n", + controller, def->data.vds.connection_id); + } - virBufferAsprintf(buffer, "ethernet%d.dvs.connectionId = \"%lld\"\n", - controller, def->data.vds.connection_id); break; } diff --git a/tests/vmx2xmldata/ethernet-vds-no-portid.vmx b/tests/vmx2xmldata/ethernet-vds-no-params.vmx similarity index 76% rename from tests/vmx2xmldata/ethernet-vds-no-portid.vmx rename to tests/vmx2xmldata/ethernet-vds-no-params.vmx index 7761accb3abc..90afbdac30cd 100644 --- a/tests/vmx2xmldata/ethernet-vds-no-portid.vmx +++ b/tests/vmx2xmldata/ethernet-vds-no-params.vmx @@ -5,6 +5,4 @@ ethernet0.virtualDev = "e1000e" ethernet0.addressType = "vpx" ethernet0.generatedAddress = "00:50:56:87:65:43" ethernet0.dvs.switchId = "50 34 26 b2 94 e9 3b 16-1d 68 87 bf ff 4a 54 40" -ethernet0.dvs.portgroupId = "dvportgroup-1285" -ethernet0.dvs.connectionId = "408217997" displayName = "test" diff --git a/tests/vmx2xmldata/ethernet-vds-no-portid.xml b/tests/vmx2xmldata/ethernet-vds-no-params.xml similarity index 82% rename from tests/vmx2xmldata/ethernet-vds-no-portid.xml rename to tests/vmx2xmldata/ethernet-vds-no-params.xml index 60fd9c99feb9..0011ba471a50 100644 --- a/tests/vmx2xmldata/ethernet-vds-no-portid.xml +++ b/tests/vmx2xmldata/ethernet-vds-no-params.xml @@ -14,7 +14,7 @@ <devices> <interface type='vds'> <mac address='00:50:56:87:65:43' type='generated'/> - <source switchid='503426b2-94e9-3b16-1d68-87bfff4a5440' portid='0' portgroupid='dvportgroup-1285' connectionid='408217997'/> + <source switchid='503426b2-94e9-3b16-1d68-87bfff4a5440'/> <model type='e1000e'/> </interface> <video> -- 2.45.1

2 1

[PATCH] test_driver: support VIR_DOMAIN_AFFECT_LIVE in testUpdateDeviceFlags()
by John Levon 04 Jul '24

04 Jul '24

Pick up some more of the qemu_driver.c code so this function supports both CONFIG and LIVE updates. Note that qemuDomainUpdateDeviceFlags() passed vm->def to virDomainDeviceDefParse() for the VIR_DOMAIN_AFFECT_CONFIG case, which is technically incorrect; in the test driver code we'll fix this. Signed-off-by: John Levon <john.levon(a)nutanix.com> --- src/test/test_driver.c | 54 +++++++++++++++++++++++++++++------------- 1 file changed, 37 insertions(+), 17 deletions(-) diff --git a/src/test/test_driver.c b/src/test/test_driver.c index 712bb20563..da682da9ad 100644 --- a/src/test/test_driver.c +++ b/src/test/test_driver.c @@ -10239,10 +10239,10 @@ testDomainAttachDevice(virDomainPtr domain, const char *xml) static int -testDomainUpdateDeviceConfig(virDomainDef *vmdef, - virDomainDeviceDef *dev, - unsigned int parse_flags, - virDomainXMLOption *xmlopt) +testDomainUpdateDevice(virDomainDef *vmdef, + virDomainDeviceDef *dev, + unsigned int parse_flags, + virDomainXMLOption *xmlopt) { virDomainDiskDef *newDisk; virDomainDeviceDef oldDev = { .type = dev->type }; @@ -10316,12 +10316,16 @@ testDomainUpdateDeviceFlags(virDomainPtr dom, testDriver *driver = dom->conn->privateData; virDomainObj *vm = NULL; virObjectEvent *event = NULL; + virDomainDef *def = NULL; + virDomainDef *persistentDef = NULL; g_autoptr(virDomainDef) vmdef = NULL; - g_autoptr(virDomainDeviceDef) dev = NULL; + g_autoptr(virDomainDeviceDef) dev_live = NULL; + g_autoptr(virDomainDeviceDef) dev_config = NULL; int ret = -1; unsigned int parse_flags = 0; - virCheckFlags(VIR_DOMAIN_AFFECT_CONFIG, -1); + virCheckFlags(VIR_DOMAIN_AFFECT_LIVE | + VIR_DOMAIN_AFFECT_CONFIG, -1); if (!(vm = testDomObjFromDomain(dom))) goto cleanup; @@ -10337,9 +10341,24 @@ testDomainUpdateDeviceFlags(virDomainPtr dom, parse_flags |= VIR_DOMAIN_DEF_PARSE_INACTIVE; } - if (!(dev = virDomainDeviceDefParse(xml, vm->def, driver->xmlopt, - NULL, parse_flags))) { - goto endjob; + /* + * NB: this has diverged from qemuDomainUpdateDeviceFlags(), which uses + * vm->def in both cases; technically the qemu driver should do the same. + */ + if (virDomainObjGetDefs(vm, flags, &def, &persistentDef) < 0) + goto cleanup; + + if (def) { + if (!(dev_live = virDomainDeviceDefParse(xml, def, driver->xmlopt, + NULL, parse_flags))) + goto endjob; + } + + if (persistentDef) { + if (!(dev_config = virDomainDeviceDefParse(xml, persistentDef, + driver->xmlopt, NULL, + parse_flags))) + goto endjob; } if (flags & VIR_DOMAIN_AFFECT_CONFIG) { @@ -10350,18 +10369,19 @@ testDomainUpdateDeviceFlags(virDomainPtr dom, /* virDomainDefCompatibleDevice call is delayed until we know the * device we're going to update. */ - if ((ret = testDomainUpdateDeviceConfig(vmdef, dev, - parse_flags, - driver->xmlopt)) < 0) + if ((ret = testDomainUpdateDevice(vmdef, dev_config, + parse_flags, + driver->xmlopt)) < 0) goto endjob; } if (flags & VIR_DOMAIN_AFFECT_LIVE) { - ret = -1; - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("live update of device '%1$s' is not supported"), - virDomainDeviceTypeToString(dev->type)); - goto endjob; + /* virDomainDefCompatibleDevice call is delayed until we know the + * device we're going to update. */ + if ((ret = testDomainUpdateDevice(def, dev_live, + parse_flags, + driver->xmlopt)) < 0) + goto endjob; } if (flags & VIR_DOMAIN_AFFECT_CONFIG) { -- 2.34.1

2 1

[PATCH] conf: Fix out-of-bounds write during cleanup of virDomainNumaDefNodeDistanceParseXML
by Rayhan Faizel 04 Jul '24

04 Jul '24

mem_nodes[i].ndistances is written outside the loop causing an out-of-bounds write leading to heap corruption. While we are at it, the entire cleanup portion can be removed as it can be handled in virDomainNumaFree. One instance of VIR_FREE is also removed and replaced with g_autofree. This patch also adds a testcase which would be picked up by ASAN, if this portion regresses. Fixes: 742494eed8dbdde8b1d05a306032334e6226beea Signed-off-by: Rayhan Faizel <rayhan.faizel(a)gmail.com> --- src/conf/numa_conf.c | 30 ++++++------------- ...ance-nonexistent-sibling.x86_64-latest.err | 1 + .../cpu-numa-distance-nonexistent-sibling.xml | 29 ++++++++++++++++++ tests/qemuxmlconftest.c | 1 + 4 files changed, 40 insertions(+), 21 deletions(-) create mode 100644 tests/qemuxmlconfdata/cpu-numa-distance-nonexistent-sibling.x86_64-latest.err create mode 100644 tests/qemuxmlconfdata/cpu-numa-distance-nonexistent-sibling.xml diff --git a/src/conf/numa_conf.c b/src/conf/numa_conf.c index d8120de6d2..0a0e2911f7 100644 --- a/src/conf/numa_conf.c +++ b/src/conf/numa_conf.c @@ -343,8 +343,7 @@ virDomainNumaFree(virDomainNuma *numa) virBitmapFree(numa->mem_nodes[i].cpumask); virBitmapFree(numa->mem_nodes[i].nodeset); - if (numa->mem_nodes[i].ndistances > 0) - g_free(numa->mem_nodes[i].distances); + g_free(numa->mem_nodes[i].distances); g_free(numa->mem_nodes[i].caches); } @@ -685,9 +684,8 @@ virDomainNumaDefNodeDistanceParseXML(virDomainNuma *def, xmlXPathContextPtr ctxt, unsigned int cur_cell) { - int ret = -1; int sibling; - xmlNodePtr *nodes = NULL; + g_autofree xmlNodePtr *nodes = NULL; size_t i, ndistances = def->nmem_nodes; if (ndistances == 0) @@ -698,12 +696,12 @@ virDomainNumaDefNodeDistanceParseXML(virDomainNuma *def, return 0; if ((sibling = virXPathNodeSet("./distances[1]/sibling", ctxt, &nodes)) < 0) - goto cleanup; + return -1; if (sibling == 0) { virReportError(VIR_ERR_XML_ERROR, "%s", _("NUMA distances defined without siblings")); - goto cleanup; + return -1; } for (i = 0; i < sibling; i++) { @@ -713,19 +711,19 @@ virDomainNumaDefNodeDistanceParseXML(virDomainNuma *def, if (virXMLPropUInt(nodes[i], "id", 10, VIR_XML_PROP_REQUIRED, &sibling_id) < 0) - goto cleanup; + return -1; /* The "id" needs to be within numa/cell range */ if (sibling_id >= ndistances) { virReportError(VIR_ERR_XML_ERROR, _("'sibling_id %1$d' does not refer to a valid cell within NUMA 'cell id %2$d'"), sibling_id, cur_cell); - goto cleanup; + return -1; } if (virXMLPropUInt(nodes[i], "value", 10, VIR_XML_PROP_REQUIRED, &sibling_value) < 0) - goto cleanup; + return -1; /* Assure LOCAL_DISTANCE <= "value" <= UNREACHABLE * and correct LOCAL_DISTANCE setting if such applies. @@ -739,7 +737,7 @@ virDomainNumaDefNodeDistanceParseXML(virDomainNuma *def, virReportError(VIR_ERR_XML_ERROR, _("'value %1$d' is invalid for 'sibling id %2$d' under NUMA 'cell id %3$d'"), sibling_value, sibling_id, cur_cell); - goto cleanup; + return -1; } /* Apply the local / remote distance */ @@ -770,17 +768,7 @@ virDomainNumaDefNodeDistanceParseXML(virDomainNuma *def, rdist[cur_cell].value = sibling_value; } - ret = 0; - - cleanup: - if (ret < 0) { - for (i = 0; i < ndistances; i++) - VIR_FREE(def->mem_nodes[i].distances); - def->mem_nodes[i].ndistances = 0; - } - VIR_FREE(nodes); - - return ret; + return 0; } diff --git a/tests/qemuxmlconfdata/cpu-numa-distance-nonexistent-sibling.x86_64-latest.err b/tests/qemuxmlconfdata/cpu-numa-distance-nonexistent-sibling.x86_64-latest.err new file mode 100644 index 0000000000..4866ff5e80 --- /dev/null +++ b/tests/qemuxmlconfdata/cpu-numa-distance-nonexistent-sibling.x86_64-latest.err @@ -0,0 +1 @@ +XML error: 'sibling_id 2' does not refer to a valid cell within NUMA 'cell id 1' diff --git a/tests/qemuxmlconfdata/cpu-numa-distance-nonexistent-sibling.xml b/tests/qemuxmlconfdata/cpu-numa-distance-nonexistent-sibling.xml new file mode 100644 index 0000000000..62a6c32fbe --- /dev/null +++ b/tests/qemuxmlconfdata/cpu-numa-distance-nonexistent-sibling.xml @@ -0,0 +1,29 @@ +<domain type='qemu'> + <name>QEMUGuest1</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory unit='KiB'>219100</memory> + <currentMemory unit='KiB'>219100</currentMemory> + <vcpu placement='static'>16</vcpu> + <os> + <type arch='x86_64' machine='pc'>hvm</type> + <boot dev='network'/> + </os> + <cpu> + <topology sockets='2' dies='1' cores='4' threads='2'/> + <numa> + <cell id='1' cpus='8-15' memory='109550' unit='KiB'> + <distances> + <sibling id='2' value='10'/> + </distances> + </cell> + <cell id='0' cpus='0-7' memory='109550' unit='KiB'/> + </numa> + </cpu> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>destroy</on_crash> + <devices> + <emulator>/usr/bin/qemu-system-x86_64</emulator> + </devices> +</domain> diff --git a/tests/qemuxmlconftest.c b/tests/qemuxmlconftest.c index 4a711fceeb..af49934c33 100644 --- a/tests/qemuxmlconftest.c +++ b/tests/qemuxmlconftest.c @@ -2160,6 +2160,7 @@ mymain(void) DO_TEST_CAPS_LATEST_PARSE_ERROR("cpu-numa3"); DO_TEST_CAPS_LATEST("cpu-numa-disjoint"); DO_TEST_CAPS_LATEST("cpu-numa-memshared"); + DO_TEST_CAPS_LATEST_PARSE_ERROR("cpu-numa-distance-nonexistent-sibling"); /* host-model cpu expansion depends on the cpu reported by qemu and thus * we invoke it for all real capability dumps we have */ -- 2.34.1

2 1

[PATCH 0/2] qemu: support logging config for swtpm
by Daniel P. Berrangé 04 Jul '24

04 Jul '24

When debugging guest problems with TPMs it is helpful to be able to have full swtpm logging. This isn't possible currently and manually restarting the swtpm process of a running guest is disruptive. Daniel P. Berrangé (2): conf: add support for 'debug' parameter on TPM emulator qemu: set swtpm log level parameter docs/formatdomain.rst | 6 ++++-- src/conf/domain_conf.c | 7 +++++++ src/conf/domain_conf.h | 1 + src/conf/schemas/domaincommon.rng | 5 +++++ src/qemu/qemu_tpm.c | 6 +++++- tests/qemuxmlconfdata/tpm-emulator-tpm2.xml | 2 +- 6 files changed, 23 insertions(+), 4 deletions(-) -- 2.45.1

2 3

[PATCH v2 0/8] ch: support restore with network devices
by Purna Pavan Chandra 04 Jul '24

04 Jul '24

Current ch driver supports restore only for domains without any network configuration defined. This was because libvirt explicitly passes network fds and CH did not had support to restore with new net FDS. This support has been added recently, https://github.com/cloud-hypervisor/cloud-hypervisor/pull/6402 The changes in this patch series includes moving to socket communication for restore api, create new net fds and pass them via SCM_RIGHTS to CH. New changes in v2: * Reword of few commints * Add version checks in save/restore validations * Add use_timeout in chSocketRecv * Address Praveen Paladugu's comments v1: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/PT77… Purna Pavan Chandra (8): ch: report response message instead of just code ch: Pass net ids explicitly during vm creation ch: refactor chProcessAddNetworkDevices ch: support poll with -1 in chSocketRecv ch: use monitor socket fd to send restore request ch: refactor virCHMonitorSaveVM ch: support restore with net devices ch: kill CH process if restore fails src/ch/ch_capabilities.c | 6 + src/ch/ch_capabilities.h | 1 + src/ch/ch_driver.c | 29 +++-- src/ch/ch_monitor.c | 62 +++++++---- src/ch/ch_monitor.h | 6 +- src/ch/ch_process.c | 233 +++++++++++++++++++++++++++++++-------- 6 files changed, 254 insertions(+), 83 deletions(-) -- 2.34.1

1 8