The aim of this function is to return whether domain definition and/or memory device that user intents to hotplug needs a private path inside cfg->memoryBackingDir. The rule for the memory device that's being hotplug includes checking whether corresponding guest NUMA node needs memoryBackingDir. Well, while the rationale behind makes sense it is not necessary to check for that really - just a few lines above every guest NUMA node was checked exactly for that. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- src/qemu/qemu_process.c | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 6a223e18c2..066f153703 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -3930,13 +3930,24 @@ qemuProcessNeedMemoryBackingPath(virDomainDefPtr def, return true; } - if (mem && - mem->model == VIR_DOMAIN_MEMORY_MODEL_DIMM && - (mem->access != VIR_DOMAIN_MEMORY_ACCESS_DEFAULT || - (mem->targetNode >= 0 && - virDomainNumaGetNodeMemoryAccessMode(def->numa, mem->targetNode) - != VIR_DOMAIN_MEMORY_ACCESS_DEFAULT))) - return true; + if (mem) { + switch (mem->model) { + case VIR_DOMAIN_MEMORY_MODEL_DIMM: + if (mem->access != VIR_DOMAIN_MEMORY_ACCESS_DEFAULT) { + /* No need to check for access mode on the target node, + * it was checked for in the previous loop. */ + return true; + } + + case VIR_DOMAIN_MEMORY_MODEL_NONE: + case VIR_DOMAIN_MEMORY_MODEL_NVDIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: + case VIR_DOMAIN_MEMORY_MODEL_LAST: + /* Backed by user provided path. Not stored in memory + * backing dir anyway. */ + break; + } + } return false; } -- 2.26.2

The virtio-mem is paravirtualized mechanism of adding/removing memory to/from a VM. A virtio-mem-pci device is split into blocks of equal size which are then exposed (all or only a requested portion of them) to the guest kernel to use as regular memory. Therefore, the device has two important attributes: 1) block-size, which defines the size of a block 2) requested-size, which defines how much memory (in bytes) is the device requested to expose to the guest. The 'block-size' is configured on command line and immutable throughout device's lifetime. The 'requested-size' can be set on the command line too, but also is adjustable via monitor. In fact, that is how management software places its requests to change the memory allocation. If it wants to give more memory to the guest it changes 'requested-size' to a bigger value, and if it wants to shrink guest memory it changes the 'requested-size' to a smaller value. Note, value of zero means that guest should release all memory offered by the device. Of course, guest has to cooperate. Therefore, there is a third attribute 'size' which is read only and reflects how much memory the guest still has. This can be different to 'requested-size', obviously. Because of name clash, I've named it 'actualsize' and it is dealt with in future commits (it is a runtime information anyway). In the backend, memory for virtio-mem is backed by usual objects: memory-backend-{ram,file,memfd} and their size puts the cap on the amount of memory that a virtio-mem device can offer to a guest. But we are already able to express this info using <size/> under <target/>. Therefore, we need only two more elements to cover 'block-size' and 'requested-size' attributes. This is the XML I've came up with: <memory model='virtio-mem'> <source> <nodemask>1-3</nodemask> <pagesize unit='KiB'>2048</pagesize> </source> <target> <size unit='KiB'>2097152</size> <node>0</node> <block unit='KiB'>2048</block> <requested unit='KiB'>1048576</requested> </target> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </memory> I hope by now it is obvious that: 1) 'requested-size' must be an integer multiple of 'block-size', and 2) virtio-mem-pci device goes onto PCI bus and thus needs PCI address. Then there is a limitation that the minimal 'block-size' is transparent huge page size (I'll leave this without explanation). Since now we have (possibly) two or more devices that allow memory inflation/deflation and accounting for all of them (and thus keeping <currentMemory/> updated) might be hard. Therefore, I'm deliberately forbidding memballoon. It's okay - virtio-mem is superior to memballoon anyway. We can always reevaluate later. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- docs/formatdomain.rst | 51 +++++++++++++-- docs/schemas/domaincommon.rng | 11 ++++ src/conf/domain_conf.c | 53 ++++++++++++++- src/conf/domain_conf.h | 3 + src/conf/domain_validate.c | 39 +++++++++++ src/qemu/qemu_alias.c | 1 + src/qemu/qemu_command.c | 1 + src/qemu/qemu_domain.c | 27 +++++++- src/qemu/qemu_domain_address.c | 38 ++++++++--- src/qemu/qemu_process.c | 2 + src/qemu/qemu_validate.c | 22 ++++++- src/security/security_apparmor.c | 1 + src/security/security_dac.c | 2 + src/security/security_selinux.c | 2 + .../memory-hotplug-virtio-mem.xml | 64 +++++++++++++++++++ ...emory-hotplug-virtio-mem.x86_64-latest.xml | 1 + tests/qemuxml2xmltest.c | 1 + 17 files changed, 298 insertions(+), 21 deletions(-) create mode 100644 tests/qemuxml2argvdata/memory-hotplug-virtio-mem.xml create mode 120000 tests/qemuxml2xmloutdata/memory-hotplug-virtio-mem.x86_64-latest.xml diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 2587106191..05d359a32d 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -6924,6 +6924,8 @@ appropriate, so there is no need to explicitly add this element in the guest XML unless a specific PCI slot needs to be assigned. :since:`Since 0.8.3, Xen, QEMU and KVM only` Additionally, :since:`since 0.8.4` , if the memballoon device needs to be explicitly disabled, ``model='none'`` may be used. +Using memballoon among with ``virtio-mem`` memory device is unsupported. In +that case, setting model to anything else then ``none`` results in an error. Example: automatically added device with KVM @@ -6935,6 +6937,16 @@ Example: automatically added device with KVM </devices> ... +Example: automatically added device with QEMU/KVM and a ``virtio-mem`` device: + +:: + + ... + <devices> + <memballoon model='none'/> + </devices> + ... + Example: manually added device with static PCI slot 2 requested :: @@ -7374,6 +7386,18 @@ Example: usage of the memory devices <size unit='KiB'>524288</size> </target> </memory> + <memory model='virtio-mem'> + <source> + <nodemask>1-3</nodemask> + <pagesize unit='KiB'>2048</pagesize> + </source> + <target> + <size unit='KiB'>2097152</size> + <node>0</node> + <block unit='KiB'>2048</block> + <requested unit='KiB'>1048576</requested> + </target> + </memory> </devices> ... @@ -7381,7 +7405,9 @@ Example: usage of the memory devices Provide ``dimm`` to add a virtual DIMM module to the guest. :since:`Since 1.2.14` Provide ``nvdimm`` model that adds a Non-Volatile DIMM module. :since:`Since 3.2.0` Provide ``virtio-pmem`` model to add a paravirtualized - persistent memory device. :since:`Since 7.1.0` + persistent memory device. :since:`Since 7.1.0` Provide ``virtio-mem`` model + to add paravirtualized memory device. :since:`Since 7.1.0` Please note that + using ``virtio-mem`` with memory balloon is unsupported. ``access`` An optional attribute ``access`` ( :since:`since 3.2.0` ) that provides @@ -7404,10 +7430,11 @@ Example: usage of the memory devices allowed only for ``model='nvdimm'`` for pSeries guests. :since:`Since 6.2.0` ``source`` - For model ``dimm`` this element is optional and allows to fine tune the - source of the memory used for the given memory device. If the element is not - provided defaults configured via ``numatune`` are used. If ``dimm`` is - provided, then the following optional elements can be provided as well: + For model ``dimm`` and model ``virtio-mem`` this element is optional and + allows to fine tune the source of the memory used for the given memory + device. If the element is not provided defaults configured via ``numatune`` + are used. If the element is provided, then the following optional elements + can be provided: ``pagesize`` This element can be used to override the default host page size used for @@ -7446,7 +7473,8 @@ Example: usage of the memory devices added memory from the perspective of the guest. The mandatory ``size`` subelement configures the size of the added memory as - a scaled integer. + a scaled integer. For ``virtio-mem`` this represents the maximum possible + size exposed to the guest. The ``node`` subelement configures the guest NUMA node to attach the memory to. The element shall be used only if the guest has NUMA nodes configured. @@ -7473,6 +7501,17 @@ Example: usage of the memory devices so other backend types should use the ``readonly`` element. :since:`Since 5.0.0` + ``block`` + For ``virtio-mem`` only. + The size of an individual block, granularity of division of memory module. + Must be power of two and at least equal to size of a transparent hugepage + (2MiB on x84_64). The default is hypervisor dependent. + + ``requested`` + For ``virtio-mem`` only. + The total size exposed to the guest. Must respect ``block`` granularity + and be smaller or equal to ``size``. + :anchor:`<a id="elementsIommu"/>` IOMMU devices diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index e6de934456..1f672756bd 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -6056,6 +6056,7 @@ <value>dimm</value> <value>nvdimm</value> <value>virtio-pmem</value> + <value>virtio-mem</value> </choice> </attribute> <optional> @@ -6140,6 +6141,16 @@ <ref name="unsignedInt"/> </element> </optional> + <optional> + <element name="block"> + <ref name="scaledInteger"/> + </element> + </optional> + <optional> + <element name="requested"> + <ref name="scaledInteger"/> + </element> + </optional> <optional> <element name="label"> <element name="size"> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index b731744f04..f4c08ef43d 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -1329,6 +1329,7 @@ VIR_ENUM_IMPL(virDomainMemoryModel, "dimm", "nvdimm", "virtio-pmem", + "virtio-mem", ); VIR_ENUM_IMPL(virDomainShmemModel, @@ -5369,6 +5370,7 @@ virDomainMemoryDefPostParse(virDomainMemoryDefPtr mem, } break; + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: case VIR_DOMAIN_MEMORY_MODEL_DIMM: case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_LAST: @@ -15392,6 +15394,7 @@ virDomainMemorySourceDefParseXML(xmlNodePtr node, switch (def->model) { case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: if (virDomainParseMemory("./pagesize", "./pagesize/@unit", ctxt, &def->pagesize, false, false) < 0) return -1; @@ -15458,7 +15461,8 @@ virDomainMemoryTargetDefParseXML(xmlNodePtr node, &def->size, true, false) < 0) return -1; - if (def->model == VIR_DOMAIN_MEMORY_MODEL_NVDIMM) { + switch (def->model) { + case VIR_DOMAIN_MEMORY_MODEL_NVDIMM: if (virDomainParseMemory("./label/size", "./label/size/@unit", ctxt, &def->labelsize, false, false) < 0) return -1; @@ -15477,6 +15481,23 @@ virDomainMemoryTargetDefParseXML(xmlNodePtr node, if (virXPathBoolean("boolean(./readonly)", ctxt)) def->readonly = true; + break; + + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: + if (virDomainParseMemory("./block", "./block/@unit", ctxt, + &def->blocksize, false, false) < 0) + return -1; + + if (virDomainParseMemory("./requested", "./requested/@unit", ctxt, + &def->requestedsize, false, false) < 0) + return -1; + break; + + case VIR_DOMAIN_MEMORY_MODEL_NONE: + case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: + case VIR_DOMAIN_MEMORY_MODEL_LAST: + break; } return 0; @@ -17294,11 +17315,14 @@ virDomainMemoryFindByDefInternal(virDomainDefPtr def, /* target info -> always present */ if (tmp->model != mem->model || tmp->targetNode != mem->targetNode || - tmp->size != mem->size) + tmp->size != mem->size || + tmp->blocksize != mem->blocksize || + tmp->requestedsize != mem->requestedsize) continue; switch (mem->model) { case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: /* source stuff -> match with device */ if (tmp->pagesize != mem->pagesize) continue; @@ -22873,6 +22897,22 @@ virDomainMemoryDefCheckABIStability(virDomainMemoryDefPtr src, return false; } + if (src->blocksize != dst->blocksize) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Target memory device block size '%llu' doesn't match " + "source memory device block size '%llu'"), + dst->blocksize, src->blocksize); + return false; + } + + if (src->requestedsize != dst->requestedsize) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Target memory device requested size '%llu' doesn't match " + "source memory device requested size '%llu'"), + dst->requestedsize, src->requestedsize); + return false; + } + if (src->model == VIR_DOMAIN_MEMORY_MODEL_NVDIMM) { if (src->labelsize != dst->labelsize) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, @@ -26629,6 +26669,7 @@ virDomainMemorySourceDefFormat(virBufferPtr buf, switch (def->model) { case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: if (def->sourceNodes) { if (!(bitmap = virBitmapFormat(def->sourceNodes))) return -1; @@ -26685,6 +26726,14 @@ virDomainMemoryTargetDefFormat(virBufferPtr buf, if (def->readonly) virBufferAddLit(&childBuf, "<readonly/>\n"); + if (def->blocksize) { + virBufferAsprintf(&childBuf, "<block unit='KiB'>%llu</block>\n", + def->blocksize); + + virBufferAsprintf(&childBuf, "<requested unit='KiB'>%llu</requested>\n", + def->requestedsize); + } + virXMLFormatElement(buf, "target", NULL, &childBuf); } diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 930eed60de..14d238df4b 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -2312,6 +2312,7 @@ typedef enum { VIR_DOMAIN_MEMORY_MODEL_DIMM, /* dimm hotpluggable memory device */ VIR_DOMAIN_MEMORY_MODEL_NVDIMM, /* nvdimm memory device */ VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM, /* virtio-pmem memory device */ + VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM, /* virtio-mem memory device */ VIR_DOMAIN_MEMORY_MODEL_LAST } virDomainMemoryModel; @@ -2332,6 +2333,8 @@ struct _virDomainMemoryDef { int targetNode; unsigned long long size; /* kibibytes */ unsigned long long labelsize; /* kibibytes; valid only for NVDIMM */ + unsigned long long blocksize; /* kibibytes; valid only for VIRTIO_MEM */ + unsigned long long requestedsize; /* kibibytes; valid only for VIRTIO_MEM */ bool readonly; /* valid only for NVDIMM */ /* required for QEMU NVDIMM ppc64 support */ diff --git a/src/conf/domain_validate.c b/src/conf/domain_validate.c index b47ecba86b..7ab91c8f2f 100644 --- a/src/conf/domain_validate.c +++ b/src/conf/domain_validate.c @@ -26,6 +26,7 @@ #include "virconftypes.h" #include "virlog.h" #include "virutil.h" +#include "virhostmem.h" #define VIR_FROM_THIS VIR_FROM_DOMAIN @@ -1612,6 +1613,8 @@ static int virDomainMemoryDefValidate(const virDomainMemoryDef *mem, const virDomainDef *def) { + unsigned long long thpSize; + switch (mem->model) { case VIR_DOMAIN_MEMORY_MODEL_NVDIMM: if (!mem->nvdimmPath) { @@ -1665,6 +1668,42 @@ virDomainMemoryDefValidate(const virDomainMemoryDef *mem, _("virtio-pmem does not support NUMA nodes")); return -1; } + break; + + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: + if (mem->requestedsize > mem->size) { + virReportError(VIR_ERR_XML_DETAIL, "%s", + _("requested size must be smaller than @size")); + return -1; + } + + if (!VIR_IS_POW2(mem->blocksize)) { + virReportError(VIR_ERR_XML_DETAIL, "%s", + _("block size must be a power of two")); + return -1; + } + + if (virHostMemGetTHPSize(&thpSize) < 0) { + /* We failed to get THP size, fall back to a sane default. On + * almost every architecture the size will be 2MiB, except for some + * funky arches like sparc and m68k. Use 2MiB and refine later if + * somebody complains. */ + thpSize = 2048; + } + + if (mem->blocksize < thpSize) { + virReportError(VIR_ERR_XML_DETAIL, + _("block size too small, must be at least %lluKiB"), + thpSize); + return -1; + } + + if (mem->requestedsize % mem->blocksize != 0) { + virReportError(VIR_ERR_XML_DETAIL, "%s", + _("requested size must be an integer multiple of block size")); + return -1; + } + break; case VIR_DOMAIN_MEMORY_MODEL_DIMM: break; diff --git a/src/qemu/qemu_alias.c b/src/qemu/qemu_alias.c index 1c6f04c0ba..1a330829a3 100644 --- a/src/qemu/qemu_alias.c +++ b/src/qemu/qemu_alias.c @@ -522,6 +522,7 @@ qemuAssignDeviceMemoryAlias(virDomainDefPtr def, case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: prefix = "virtiopmem"; break; + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_LAST: default: diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index d801018aa2..ac9a8c9f6a 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -3355,6 +3355,7 @@ qemuBuildMemoryDeviceStr(const virDomainDef *def, device = "virtio-pmem-pci"; break; + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_LAST: default: diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 541d592bbe..26b9b5583e 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -3679,10 +3679,21 @@ qemuDomainDefAddDefaultDevices(virDomainDefPtr def, } if (addDefaultMemballoon && !def->memballoon) { - virDomainMemballoonDefPtr memballoon; - memballoon = g_new0(virDomainMemballoonDef, 1); + virDomainMemballoonDefPtr memballoon = g_new0(virDomainMemballoonDef, 1); + size_t i; - memballoon->model = VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO; + /* To simplify virtio-mem implementation, memballoon has to be turned + * off if domain has a virtio-mem device. See + * qemuValidateDomainDeviceDefMemory() for more details. */ + for (i = 0; i < def->nmems; i++) { + if (def->mems[i]->model == VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM) + break; + } + + if (i == def->nmems) + memballoon->model = VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO; + else + memballoon->model = VIR_DOMAIN_MEMBALLOON_MODEL_NONE; def->memballoon = memballoon; } @@ -8775,6 +8786,16 @@ qemuDomainDefValidateMemoryHotplugDevice(const virDomainMemoryDef *mem, needsNuma = false; break; + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: + if (mem->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI && + mem->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("only 'pci' addresses are supported for the %s device"), + virDomainMemoryModelTypeToString(mem->model)); + return -1; + } + break; + case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_LAST: return -1; diff --git a/src/qemu/qemu_domain_address.c b/src/qemu/qemu_domain_address.c index 68dbf9e95b..64c57f1cad 100644 --- a/src/qemu/qemu_domain_address.c +++ b/src/qemu/qemu_domain_address.c @@ -380,9 +380,18 @@ qemuDomainPrimeVirtioDeviceAddresses(virDomainDefPtr def, } for (i = 0; i < def->nmems; i++) { - if (def->mems[i]->model == VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM && - def->mems[i]->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) - def->mems[i]->info.type = type; + switch (def->mems[i]->model) { + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: + if (def->mems[i]->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) + def->mems[i]->info.type = type; + break; + case VIR_DOMAIN_MEMORY_MODEL_NONE: + case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_NVDIMM: + case VIR_DOMAIN_MEMORY_MODEL_LAST: + break; + } } if (type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW) { @@ -1011,6 +1020,7 @@ qemuDomainDeviceCalculatePCIConnectFlags(virDomainDeviceDefPtr dev, case VIR_DOMAIN_DEVICE_MEMORY: switch (dev->data.memory->model) { case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: return virtioFlags; case VIR_DOMAIN_MEMORY_MODEL_NONE: @@ -2438,12 +2448,19 @@ qemuDomainAssignDevicePCISlots(virDomainDefPtr def, for (i = 0; i < def->nmems; i++) { virDomainMemoryDefPtr mem = def->mems[i]; - if (mem->model != VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM || - !virDeviceInfoPCIAddressIsWanted(&mem->info)) - continue; - - if (qemuDomainPCIAddressReserveNextAddr(addrs, &mem->info) < 0) - return -1; + switch (mem->model) { + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: + if (virDeviceInfoPCIAddressIsWanted(&mem->info) && + qemuDomainPCIAddressReserveNextAddr(addrs, &mem->info) < 0) + return -1; + break; + case VIR_DOMAIN_MEMORY_MODEL_NONE: + case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_NVDIMM: + case VIR_DOMAIN_MEMORY_MODEL_LAST: + break; + } } return 0; @@ -3110,6 +3127,7 @@ qemuDomainAssignMemoryDeviceSlot(virQEMUDriverPtr driver, break; case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: return qemuDomainEnsurePCIAddress(vm, &dev, driver); break; @@ -3135,6 +3153,7 @@ qemuDomainReleaseMemoryDeviceSlot(virDomainObjPtr vm, break; case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: qemuDomainReleaseDeviceAddress(vm, &mem->info); break; @@ -3169,6 +3188,7 @@ qemuDomainAssignMemorySlots(virDomainDefPtr def) break; case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: /* handled in qemuDomainAssignPCIAddresses() */ break; case VIR_DOMAIN_MEMORY_MODEL_NONE: diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 066f153703..ebf42259b5 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -3869,6 +3869,7 @@ qemuProcessDomainMemoryDefNeedHugepagesPath(const virDomainMemoryDef *mem, { switch (mem->model) { case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: return mem->pagesize && mem->pagesize != system_pagesize; @@ -3933,6 +3934,7 @@ qemuProcessNeedMemoryBackingPath(virDomainDefPtr def, if (mem) { switch (mem->model) { case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: if (mem->access != VIR_DOMAIN_MEMORY_ACCESS_DEFAULT) { /* No need to check for access mode on the target node, * it was checked for in the previous loop. */ diff --git a/src/qemu/qemu_validate.c b/src/qemu/qemu_validate.c index 2541ae856a..6150083251 100644 --- a/src/qemu/qemu_validate.c +++ b/src/qemu/qemu_validate.c @@ -4653,6 +4653,7 @@ qemuValidateDomainDeviceDefHub(virDomainHubDefPtr hub, static int qemuValidateDomainDeviceDefMemory(virDomainMemoryDefPtr mem, + const virDomainDef *def, virQEMUCapsPtr qemuCaps) { switch (mem->model) { @@ -4688,6 +4689,25 @@ qemuValidateDomainDeviceDefMemory(virDomainMemoryDefPtr mem, } break; + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: + /* Accounting balloon and virtio-mem is hard. We have plenty of APIs + * which take balloon from QEMU and report it to users. We would have + * to change all that and account for virtio-mem actual size. Also, + * virtio-mem is supposed to be replacement for balloon. Disable + * coexistence of these two for now. */ + if (virDomainDefHasMemballoon(def)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("virtio-mem is not supported with memory balloon")); + return -1; + } + + if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIRTIO_MEM_PCI)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("virtio-mem isn't supported by this QEMU binary")); + return -1; + } + break; + case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_LAST: break; @@ -4842,7 +4862,7 @@ qemuValidateDomainDeviceDef(const virDomainDeviceDef *dev, break; case VIR_DOMAIN_DEVICE_MEMORY: - ret = qemuValidateDomainDeviceDefMemory(dev->data.memory, qemuCaps); + ret = qemuValidateDomainDeviceDefMemory(dev->data.memory, def, qemuCaps); break; case VIR_DOMAIN_DEVICE_SHMEM: diff --git a/src/security/security_apparmor.c b/src/security/security_apparmor.c index e6ecb513b1..5aeaf975a1 100644 --- a/src/security/security_apparmor.c +++ b/src/security/security_apparmor.c @@ -693,6 +693,7 @@ AppArmorSetMemoryLabel(virSecurityManagerPtr mgr, return reload_profile(mgr, def, mem->nvdimmPath, true); case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: case VIR_DOMAIN_MEMORY_MODEL_LAST: break; } diff --git a/src/security/security_dac.c b/src/security/security_dac.c index 00eeae0d27..5ae645f25b 100644 --- a/src/security/security_dac.c +++ b/src/security/security_dac.c @@ -1871,6 +1871,7 @@ virSecurityDACRestoreMemoryLabel(virSecurityManagerPtr mgr, break; case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: case VIR_DOMAIN_MEMORY_MODEL_LAST: case VIR_DOMAIN_MEMORY_MODEL_NONE: ret = 0; @@ -2054,6 +2055,7 @@ virSecurityDACSetMemoryLabel(virSecurityManagerPtr mgr, break; case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: case VIR_DOMAIN_MEMORY_MODEL_LAST: case VIR_DOMAIN_MEMORY_MODEL_NONE: ret = 0; diff --git a/src/security/security_selinux.c b/src/security/security_selinux.c index 1d1d9edfff..c0a3b28044 100644 --- a/src/security/security_selinux.c +++ b/src/security/security_selinux.c @@ -1584,6 +1584,7 @@ virSecuritySELinuxSetMemoryLabel(virSecurityManagerPtr mgr, case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: case VIR_DOMAIN_MEMORY_MODEL_LAST: break; } @@ -1611,6 +1612,7 @@ virSecuritySELinuxRestoreMemoryLabel(virSecurityManagerPtr mgr, break; case VIR_DOMAIN_MEMORY_MODEL_DIMM: + case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_LAST: ret = 0; diff --git a/tests/qemuxml2argvdata/memory-hotplug-virtio-mem.xml b/tests/qemuxml2argvdata/memory-hotplug-virtio-mem.xml new file mode 100644 index 0000000000..eed9e35d0d --- /dev/null +++ b/tests/qemuxml2argvdata/memory-hotplug-virtio-mem.xml @@ -0,0 +1,64 @@ +<domain type='kvm'> + <name>QEMUGuest1</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <maxMemory slots='16' unit='KiB'>1099511627776</maxMemory> + <memory unit='KiB'>8388608</memory> + <currentMemory unit='KiB'>8388608</currentMemory> + <vcpu placement='static' cpuset='0-1'>2</vcpu> + <os> + <type arch='x86_64' machine='pc'>hvm</type> + <boot dev='hd'/> + </os> + <cpu mode='custom' match='exact' check='none'> + <model fallback='forbid'>qemu64</model> + <topology sockets='2' dies='1' cores='1' threads='1'/> + <numa> + <cell id='0' cpus='0-1' memory='2095104' 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-i386</emulator> + <disk type='block' device='disk'> + <driver name='qemu' type='raw'/> + <source dev='/dev/HostVG/QEMUGuest1'/> + <target dev='hda' bus='ide'/> + <address type='drive' controller='0' bus='0' target='0' unit='0'/> + </disk> + <controller type='ide' index='0'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x1'/> + </controller> + <controller type='usb' index='0' model='piix3-uhci'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x2'/> + </controller> + <controller type='pci' index='0' model='pci-root'/> + <input type='mouse' bus='ps2'/> + <input type='keyboard' bus='ps2'/> + <memballoon model='none'/> + <memory model='virtio-mem'> + <target> + <size unit='KiB'>1048576</size> + <node>0</node> + <block unit='KiB'>2048</block> + <requested unit='KiB'>524288</requested> + </target> + <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> + </memory> + <memory model='virtio-mem'> + <source> + <nodemask>1-3</nodemask> + <pagesize unit='KiB'>2048</pagesize> + </source> + <target> + <size unit='KiB'>2097152</size> + <node>0</node> + <block unit='KiB'>2048</block> + <requested unit='KiB'>1048576</requested> + </target> + <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> + </memory> + </devices> +</domain> diff --git a/tests/qemuxml2xmloutdata/memory-hotplug-virtio-mem.x86_64-latest.xml b/tests/qemuxml2xmloutdata/memory-hotplug-virtio-mem.x86_64-latest.xml new file mode 120000 index 0000000000..a9d298129c --- /dev/null +++ b/tests/qemuxml2xmloutdata/memory-hotplug-virtio-mem.x86_64-latest.xml @@ -0,0 +1 @@ +../qemuxml2argvdata/memory-hotplug-virtio-mem.xml \ No newline at end of file diff --git a/tests/qemuxml2xmltest.c b/tests/qemuxml2xmltest.c index 5cd945f28f..3a77f8d6c0 100644 --- a/tests/qemuxml2xmltest.c +++ b/tests/qemuxml2xmltest.c @@ -1267,6 +1267,7 @@ mymain(void) QEMU_CAPS_DEVICE_NVDIMM, QEMU_CAPS_LAST); DO_TEST_CAPS_LATEST("memory-hotplug-virtio-pmem"); + DO_TEST_CAPS_LATEST("memory-hotplug-virtio-mem"); DO_TEST("net-udp", NONE); -- 2.26.2

Nothing special is happening here. All important changes were done when for 'virtio-pmem' (adjusting the code to put virtio memory on PCI bus, generating alias using qemuDomainDeviceAliasIndex(). The only bit that might look suspicious is no prealloc for virtio-mem. But if you think about it, the whole purpose of this device is to change amount of memory exposed to guest on the fly. There is no point in locking the whole backend in memory. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- src/qemu/qemu_alias.c | 9 +++- src/qemu/qemu_command.c | 12 ++++- ...mory-hotplug-virtio-mem.x86_64-latest.args | 48 +++++++++++++++++++ tests/qemuxml2argvtest.c | 1 + 4 files changed, 67 insertions(+), 3 deletions(-) create mode 100644 tests/qemuxml2argvdata/memory-hotplug-virtio-mem.x86_64-latest.args diff --git a/src/qemu/qemu_alias.c b/src/qemu/qemu_alias.c index 1a330829a3..cef02ae916 100644 --- a/src/qemu/qemu_alias.c +++ b/src/qemu/qemu_alias.c @@ -475,8 +475,11 @@ qemuDeviceMemoryGetAliasID(virDomainDefPtr def, size_t i; int maxidx = 0; - /* virtio-pmem goes onto PCI bus and thus DIMM address is not valid */ - if (!oldAlias && mem->model != VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM) + /* virtio-pmem and virtio-mem go onto PCI bus and thus DIMM address is not + * valid */ + if (!oldAlias && + mem->model != VIR_DOMAIN_MEMORY_MODEL_VIRTIO_PMEM && + mem->model != VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM) return mem->info.addr.dimm.slot; for (i = 0; i < def->nmems; i++) { @@ -523,6 +526,8 @@ qemuAssignDeviceMemoryAlias(virDomainDefPtr def, prefix = "virtiopmem"; break; case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: + prefix = "virtiomem"; + break; case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_LAST: default: diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index ac9a8c9f6a..d56ba26874 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -3127,7 +3127,9 @@ qemuBuildMemoryBackendProps(virJSONValuePtr *backendProps, } else if (useHugepage) { if (qemuGetDomainHupageMemPath(priv->driver, def, pagesize, &memPath) < 0) return -1; - prealloc = true; + /* For virtio-mem backed by hugepages we don't need prealloc. */ + if (mem->model != VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM) + prealloc = true; } else { /* We can have both pagesize and mem source. If that's the case, * prefer hugepages as those are more specific. */ @@ -3356,6 +3358,9 @@ qemuBuildMemoryDeviceStr(const virDomainDef *def, break; case VIR_DOMAIN_MEMORY_MODEL_VIRTIO_MEM: + device = "virtio-mem-pci"; + break; + case VIR_DOMAIN_MEMORY_MODEL_NONE: case VIR_DOMAIN_MEMORY_MODEL_LAST: default: @@ -3372,6 +3377,11 @@ qemuBuildMemoryDeviceStr(const virDomainDef *def, if (mem->labelsize) virBufferAsprintf(&buf, "label-size=%llu,", mem->labelsize * 1024); + if (mem->blocksize) { + virBufferAsprintf(&buf, "block-size=%llu,", mem->blocksize * 1024); + virBufferAsprintf(&buf, "requested-size=%llu,", mem->requestedsize * 1024); + } + if (mem->uuid) { char uuidstr[VIR_UUID_STRING_BUFLEN]; diff --git a/tests/qemuxml2argvdata/memory-hotplug-virtio-mem.x86_64-latest.args b/tests/qemuxml2argvdata/memory-hotplug-virtio-mem.x86_64-latest.args new file mode 100644 index 0000000000..76d8c41886 --- /dev/null +++ b/tests/qemuxml2argvdata/memory-hotplug-virtio-mem.x86_64-latest.args @@ -0,0 +1,48 @@ +LC_ALL=C \ +PATH=/bin \ +HOME=/tmp/lib/domain--1-QEMUGuest1 \ +USER=test \ +LOGNAME=test \ +XDG_DATA_HOME=/tmp/lib/domain--1-QEMUGuest1/.local/share \ +XDG_CACHE_HOME=/tmp/lib/domain--1-QEMUGuest1/.cache \ +XDG_CONFIG_HOME=/tmp/lib/domain--1-QEMUGuest1/.config \ +QEMU_AUDIO_DRV=none \ +/usr/bin/qemu-system-i386 \ +-name guest=QEMUGuest1,debug-threads=on \ +-S \ +-object secret,id=masterKey0,format=raw,\ +file=/tmp/lib/domain--1-QEMUGuest1/master-key.aes \ +-machine pc,accel=kvm,usb=off,dump-guest-core=off \ +-cpu qemu64 \ +-m size=2095104k,slots=16,maxmem=1099511627776k \ +-overcommit mem-lock=off \ +-smp 2,sockets=2,dies=1,cores=1,threads=1 \ +-object memory-backend-ram,id=ram-node0,size=2145386496 \ +-numa node,nodeid=0,cpus=0-1,memdev=ram-node0 \ +-object memory-backend-ram,id=memvirtiomem0,size=1073741824 \ +-device virtio-mem-pci,node=0,block-size=2097152,requested-size=536870912,\ +memdev=memvirtiomem0,id=virtiomem0,bus=pci.0,addr=0x2 \ +-object memory-backend-file,id=memvirtiomem1,\ +mem-path=/dev/hugepages2M/libvirt/qemu/-1-QEMUGuest1,size=2147483648,\ +host-nodes=1-3,policy=bind \ +-device virtio-mem-pci,node=0,block-size=2097152,requested-size=1073741824,\ +memdev=memvirtiomem1,id=virtiomem1,bus=pci.0,addr=0x3 \ +-uuid c7a5fdbd-edaf-9455-926a-d65c16db1809 \ +-display none \ +-no-user-config \ +-nodefaults \ +-chardev socket,id=charmonitor,fd=1729,server=on,wait=off \ +-mon chardev=charmonitor,id=monitor,mode=control \ +-rtc base=utc \ +-no-shutdown \ +-no-acpi \ +-boot strict=on \ +-device piix3-usb-uhci,id=usb,bus=pci.0,addr=0x1.0x2 \ +-blockdev '{"driver":"host_device","filename":"/dev/HostVG/QEMUGuest1",\ +"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}' \ +-blockdev '{"node-name":"libvirt-1-format","read-only":false,"driver":"raw",\ +"file":"libvirt-1-storage"}' \ +-device ide-hd,bus=ide.0,unit=0,drive=libvirt-1-format,id=ide0-0-0,bootindex=1 \ +-sandbox on,obsolete=deny,elevateprivileges=deny,spawn=deny,\ +resourcecontrol=deny \ +-msg timestamp=on diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index db438c5466..61aff38abd 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -3060,6 +3060,7 @@ mymain(void) QEMU_CAPS_DEVICE_NVDIMM, QEMU_CAPS_LAST); DO_TEST_CAPS_LATEST("memory-hotplug-virtio-pmem"); + DO_TEST_CAPS_LATEST("memory-hotplug-virtio-mem"); DO_TEST("machine-aeskeywrap-on-caps", QEMU_CAPS_AES_KEY_WRAP, -- 2.26.2

As advertised in previous commit, this event is delivered to us when virtio-mem module changes the allocation inside the guest. It comes with one attribute - size - which holds the new size of the virtio-mem (well, allocated size), in bytes. Mind you, this is not necessarily the same number as 'requested size'. It almost certainly will be when sizing the memory up, but it might not be when sizing the memory down - the guest kernel might be unable to free some blocks. This actual size is reported in the domain XML as an output element only. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- docs/formatdomain.rst | 7 ++++++ docs/schemas/domaincommon.rng | 5 +++++ src/conf/domain_conf.c | 26 ++++++++++++++++++++-- src/conf/domain_conf.h | 8 +++++++ src/libvirt_private.syms | 1 + src/qemu/qemu_domain.c | 3 +++ src/qemu/qemu_domain.h | 1 + src/qemu/qemu_driver.c | 31 ++++++++++++++++++++++++++ src/qemu/qemu_monitor.c | 24 ++++++++++++++++++++ src/qemu/qemu_monitor.h | 20 +++++++++++++++++ src/qemu/qemu_monitor_json.c | 24 ++++++++++++++++++++ src/qemu/qemu_process.c | 42 +++++++++++++++++++++++++++++++++++ 12 files changed, 190 insertions(+), 2 deletions(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 05d359a32d..ccce01fd3b 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -7396,6 +7396,7 @@ Example: usage of the memory devices <node>0</node> <block unit='KiB'>2048</block> <requested unit='KiB'>1048576</requested> + <actual unit='KiB'>524288</requested> </target> </memory> </devices> @@ -7512,6 +7513,12 @@ Example: usage of the memory devices The total size exposed to the guest. Must respect ``block`` granularity and be smaller or equal to ``size``. + ``actual`` + Active XML for ``virtio-mem`` model may contain ``actual`` element that + reflects the actual size of the corresponding virtio memory device. The + element is formatted into live XML and never parsed, i.e. it is + output-only element. + :anchor:`<a id="elementsIommu"/>` IOMMU devices diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index 1f672756bd..fc7569ba8f 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -6151,6 +6151,11 @@ <ref name="scaledInteger"/> </element> </optional> + <optional> + <element name="actual"> + <ref name="scaledInteger"/> + </element> + </optional> <optional> <element name="label"> <element name="size"> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 5f9210a112..aaeb50f2a2 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -17413,6 +17413,23 @@ virDomainMemoryFindByDeviceInfo(virDomainDefPtr def, } +virDomainMemoryDefPtr +virDomainMemoryFindByDeviceAlias(virDomainDefPtr def, + const char *alias) +{ + size_t i; + + for (i = 0; i < def->nmems; i++) { + virDomainMemoryDefPtr tmp = def->mems[i]; + + if (STREQ_NULLABLE(tmp->info.alias, alias)) + return tmp; + } + + return NULL; +} + + /** * virDomainMemoryInsert: * @@ -26746,7 +26763,8 @@ virDomainMemorySourceDefFormat(virBufferPtr buf, static void virDomainMemoryTargetDefFormat(virBufferPtr buf, - virDomainMemoryDefPtr def) + virDomainMemoryDefPtr def, + unsigned int flags) { g_auto(virBuffer) childBuf = VIR_BUFFER_INIT_CHILD(buf); @@ -26768,6 +26786,10 @@ virDomainMemoryTargetDefFormat(virBufferPtr buf, virBufferAsprintf(&childBuf, "<requested unit='KiB'>%llu</requested>\n", def->requestedsize); + if (!(flags & VIR_DOMAIN_DEF_FORMAT_INACTIVE)) { + virBufferAsprintf(&childBuf, "<actual unit='KiB'>%llu</actual>\n", + def->actualsize); + } } virXMLFormatElement(buf, "target", NULL, &childBuf); @@ -26800,7 +26822,7 @@ virDomainMemoryDefFormat(virBufferPtr buf, if (virDomainMemorySourceDefFormat(buf, def) < 0) return -1; - virDomainMemoryTargetDefFormat(buf, def); + virDomainMemoryTargetDefFormat(buf, def, flags); virDomainDeviceInfoFormat(buf, &def->info, flags); diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index af38d8aa81..8a43c9655f 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -2335,6 +2335,9 @@ struct _virDomainMemoryDef { unsigned long long labelsize; /* kibibytes; valid only for NVDIMM */ unsigned long long blocksize; /* kibibytes; valid only for VIRTIO_MEM */ unsigned long long requestedsize; /* kibibytes; valid only for VIRTIO_MEM */ + unsigned long long actualsize; /* kibibytes, valid for VIRTIO_MEM and + active domain only, only to report never + parse */ bool readonly; /* valid only for NVDIMM */ /* required for QEMU NVDIMM ppc64 support */ @@ -3627,6 +3630,11 @@ virDomainMemoryFindByDeviceInfo(virDomainDefPtr dev, virDomainDeviceInfoPtr info) ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2) G_GNUC_WARN_UNUSED_RESULT; +virDomainMemoryDefPtr +virDomainMemoryFindByDeviceAlias(virDomainDefPtr def, + const char *alias) + ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2) G_GNUC_WARN_UNUSED_RESULT; + int virDomainShmemDefInsert(virDomainDefPtr def, virDomainShmemDefPtr shmem) ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2) G_GNUC_WARN_UNUSED_RESULT; bool virDomainShmemDefEquals(virDomainShmemDefPtr src, virDomainShmemDefPtr dst) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index ec1af101a9..4f362e3c54 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -494,6 +494,7 @@ virDomainMemballoonModelTypeFromString; virDomainMemballoonModelTypeToString; virDomainMemoryDefFree; virDomainMemoryFindByDef; +virDomainMemoryFindByDeviceAlias; virDomainMemoryFindByDeviceInfo; virDomainMemoryFindInactiveByDef; virDomainMemoryInsert; diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 26b9b5583e..7de9f8c050 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -10880,6 +10880,9 @@ qemuProcessEventFree(struct qemuProcessEvent *event) case QEMU_PROCESS_EVENT_JOB_STATUS_CHANGE: virObjectUnref(event->data); break; + case QEMU_PROCESS_EVENT_MEMORY_DEVICE_SIZE_CHANGE: + qemuMonitorMemoryDeviceSizeChangeFree(event->data); + break; case QEMU_PROCESS_EVENT_PR_DISCONNECT: case QEMU_PROCESS_EVENT_LAST: break; diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h index 9a7d997d65..91d6d6ef80 100644 --- a/src/qemu/qemu_domain.h +++ b/src/qemu/qemu_domain.h @@ -441,6 +441,7 @@ typedef enum { QEMU_PROCESS_EVENT_PR_DISCONNECT, QEMU_PROCESS_EVENT_RDMA_GID_STATUS_CHANGED, QEMU_PROCESS_EVENT_GUEST_CRASHLOADED, + QEMU_PROCESS_EVENT_MEMORY_DEVICE_SIZE_CHANGE, QEMU_PROCESS_EVENT_LAST } qemuProcessEventType; diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 1a121e9278..995faf2abc 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -4279,6 +4279,34 @@ processGuestCrashloadedEvent(virQEMUDriverPtr driver, } +static void +processMemoryDeviceSizeChange(virQEMUDriverPtr driver, + virDomainObjPtr vm, + qemuMonitorMemoryDeviceSizeChangePtr info) +{ + virDomainMemoryDefPtr mem = NULL; + + if (qemuDomainObjBeginJob(driver, vm, QEMU_JOB_MODIFY) < 0) + return; + + if (!virDomainObjIsActive(vm)) { + VIR_DEBUG("Domain is not running"); + goto endjob; + } + + mem = virDomainMemoryFindByDeviceAlias(vm->def, info->devAlias); + if (!mem) { + VIR_DEBUG("Memory device '%s' not found", info->devAlias); + goto endjob; + } + + mem->actualsize = VIR_DIV_UP(info->size, 1024); + + endjob: + qemuDomainObjEndJob(driver, vm); +} + + static void qemuProcessEventHandler(void *data, void *opaque) { struct qemuProcessEvent *processEvent = data; @@ -4328,6 +4356,9 @@ static void qemuProcessEventHandler(void *data, void *opaque) case QEMU_PROCESS_EVENT_GUEST_CRASHLOADED: processGuestCrashloadedEvent(driver, vm); break; + case QEMU_PROCESS_EVENT_MEMORY_DEVICE_SIZE_CHANGE: + processMemoryDeviceSizeChange(driver, vm, processEvent->data); + break; case QEMU_PROCESS_EVENT_LAST: break; } diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c index 217870c437..a7b7a78a55 100644 --- a/src/qemu/qemu_monitor.c +++ b/src/qemu/qemu_monitor.c @@ -1437,6 +1437,20 @@ qemuMonitorEmitSpiceMigrated(qemuMonitorPtr mon) } +int +qemuMonitorEmitMemoryDeviceSizeChange(qemuMonitorPtr mon, + const char *devAlias, + unsigned long long size) +{ + int ret = -1; + VIR_DEBUG("mon=%p, devAlias='%s', size=%llu", mon, devAlias, size); + + QEMU_MONITOR_CALLBACK(mon, ret, domainMemoryDeviceSizeChange, mon->vm, devAlias, size); + + return ret; +} + + int qemuMonitorEmitMemoryFailure(qemuMonitorPtr mon, qemuMonitorEventMemoryFailurePtr mfp) @@ -4408,6 +4422,16 @@ qemuMonitorEventRdmaGidStatusFree(qemuMonitorRdmaGidStatusPtr info) } +void +qemuMonitorMemoryDeviceSizeChangeFree(qemuMonitorMemoryDeviceSizeChangePtr info) +{ + if (!info) + return; + + g_free(info->devAlias); +} + + int qemuMonitorSetWatchdogAction(qemuMonitorPtr mon, const char *action) diff --git a/src/qemu/qemu_monitor.h b/src/qemu/qemu_monitor.h index 05df73cff8..53f2aa9600 100644 --- a/src/qemu/qemu_monitor.h +++ b/src/qemu/qemu_monitor.h @@ -108,6 +108,14 @@ struct _qemuMonitorRdmaGidStatus { }; +typedef struct _qemuMonitorMemoryDeviceSizeChange qemuMonitorMemoryDeviceSizeChange; +typedef qemuMonitorMemoryDeviceSizeChange *qemuMonitorMemoryDeviceSizeChangePtr; +struct _qemuMonitorMemoryDeviceSizeChange { + char *devAlias; + unsigned long long size; +}; + + typedef enum { QEMU_MONITOR_JOB_TYPE_UNKNOWN, /* internal value, not exposed by qemu */ QEMU_MONITOR_JOB_TYPE_COMMIT, @@ -153,6 +161,7 @@ struct _qemuMonitorJobInfo { char *qemuMonitorGuestPanicEventInfoFormatMsg(qemuMonitorEventPanicInfoPtr info); void qemuMonitorEventPanicInfoFree(qemuMonitorEventPanicInfoPtr info); void qemuMonitorEventRdmaGidStatusFree(qemuMonitorRdmaGidStatusPtr info); +void qemuMonitorMemoryDeviceSizeChangeFree(qemuMonitorMemoryDeviceSizeChangePtr info); typedef void (*qemuMonitorDestroyCallback)(qemuMonitorPtr mon, virDomainObjPtr vm, @@ -374,6 +383,12 @@ typedef int (*qemuMonitorDomainMemoryFailureCallback)(qemuMonitorPtr mon, qemuMonitorEventMemoryFailurePtr mfp, void *opaque); +typedef int (*qemuMonitorDomainMemoryDeviceSizeChange)(qemuMonitorPtr mon, + virDomainObjPtr vm, + const char *alias, + unsigned long long size, + void *opaque); + typedef struct _qemuMonitorCallbacks qemuMonitorCallbacks; typedef qemuMonitorCallbacks *qemuMonitorCallbacksPtr; struct _qemuMonitorCallbacks { @@ -411,6 +426,7 @@ struct _qemuMonitorCallbacks { qemuMonitorDomainRdmaGidStatusChangedCallback domainRdmaGidStatusChanged; qemuMonitorDomainGuestCrashloadedCallback domainGuestCrashloaded; qemuMonitorDomainMemoryFailureCallback domainMemoryFailure; + qemuMonitorDomainMemoryDeviceSizeChange domainMemoryDeviceSizeChange; }; qemuMonitorPtr qemuMonitorOpen(virDomainObjPtr vm, @@ -507,6 +523,10 @@ int qemuMonitorEmitSerialChange(qemuMonitorPtr mon, bool connected); int qemuMonitorEmitSpiceMigrated(qemuMonitorPtr mon); +int qemuMonitorEmitMemoryDeviceSizeChange(qemuMonitorPtr mon, + const char *devAlias, + unsigned long long size); + int qemuMonitorEmitMemoryFailure(qemuMonitorPtr mon, qemuMonitorEventMemoryFailurePtr mfp); diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 07d19b5e6d..9ce7ba52ba 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -113,6 +113,7 @@ static void qemuMonitorJSONHandleDumpCompleted(qemuMonitorPtr mon, virJSONValueP static void qemuMonitorJSONHandlePRManagerStatusChanged(qemuMonitorPtr mon, virJSONValuePtr data); static void qemuMonitorJSONHandleRdmaGidStatusChanged(qemuMonitorPtr mon, virJSONValuePtr data); static void qemuMonitorJSONHandleMemoryFailure(qemuMonitorPtr mon, virJSONValuePtr data); +static void qemuMonitorJSONHandleMemoryDeviceSizeChange(qemuMonitorPtr mon, virJSONValuePtr data); typedef struct { const char *type; @@ -133,6 +134,7 @@ static qemuEventHandler eventHandlers[] = { { "GUEST_CRASHLOADED", qemuMonitorJSONHandleGuestCrashloaded, }, { "GUEST_PANICKED", qemuMonitorJSONHandleGuestPanic, }, { "JOB_STATUS_CHANGE", qemuMonitorJSONHandleJobStatusChange, }, + { "MEMORY_DEVICE_SIZE_CHANGE", qemuMonitorJSONHandleMemoryDeviceSizeChange, }, { "MEMORY_FAILURE", qemuMonitorJSONHandleMemoryFailure, }, { "MIGRATION", qemuMonitorJSONHandleMigrationStatus, }, { "MIGRATION_PASS", qemuMonitorJSONHandleMigrationPass, }, @@ -1333,6 +1335,28 @@ qemuMonitorJSONHandleSpiceMigrated(qemuMonitorPtr mon, } +static void +qemuMonitorJSONHandleMemoryDeviceSizeChange(qemuMonitorPtr mon, + virJSONValuePtr data) +{ + const char *name; + unsigned long long size; + + if (!(name = virJSONValueObjectGetString(data, "id"))) { + VIR_WARN("missing device alias in MEMORY_DEVICE_SIZE_CHANGE event"); + return; + } + + if (virJSONValueObjectGetNumberUlong(data, "size", &size) < 0) { + VIR_WARN("missing new size for '%s' in MEMORY_DEVICE_SIZE_CHANGE event", name); + return; + } + + + qemuMonitorEmitMemoryDeviceSizeChange(mon, name, size); +} + + static void qemuMonitorJSONHandleMemoryFailure(qemuMonitorPtr mon, virJSONValuePtr data) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index ebf42259b5..4ab085841a 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -1937,6 +1937,47 @@ qemuProcessHandleMemoryFailure(qemuMonitorPtr mon G_GNUC_UNUSED, } +static int +qemuProcessHandleMemoryDeviceSizeChange(qemuMonitorPtr mon G_GNUC_UNUSED, + virDomainObjPtr vm, + const char *devAlias, + unsigned long long size, + void *opaque) +{ + virQEMUDriverPtr driver = opaque; + struct qemuProcessEvent *processEvent = NULL; + qemuMonitorMemoryDeviceSizeChangePtr info = NULL; + int ret = -1; + + virObjectLock(vm); + + VIR_DEBUG("Memory device '%s' changed size to '%llu' in domain '%s'", + devAlias, size, vm->def->name); + + info = g_new0(qemuMonitorMemoryDeviceSizeChange, 1); + info->devAlias = g_strdup(devAlias); + info->size = size; + + processEvent = g_new0(struct qemuProcessEvent, 1); + processEvent->eventType = QEMU_PROCESS_EVENT_MEMORY_DEVICE_SIZE_CHANGE; + processEvent->vm = virObjectRef(vm); + processEvent->data = g_steal_pointer(&info); + + if (virThreadPoolSendJob(driver->workerPool, 0, processEvent) < 0) { + qemuProcessEventFree(processEvent); + virObjectUnref(vm); + goto cleanup; + } + + processEvent = NULL; + ret = 0; + cleanup: + qemuProcessEventFree(processEvent); + virObjectUnlock(vm); + return ret; +} + + static qemuMonitorCallbacks monitorCallbacks = { .eofNotify = qemuProcessHandleMonitorEOF, .errorNotify = qemuProcessHandleMonitorError, @@ -1970,6 +2011,7 @@ static qemuMonitorCallbacks monitorCallbacks = { .domainRdmaGidStatusChanged = qemuProcessHandleRdmaGidStatusChanged, .domainGuestCrashloaded = qemuProcessHandleGuestCrashloaded, .domainMemoryFailure = qemuProcessHandleMemoryFailure, + .domainMemoryDeviceSizeChange = qemuProcessHandleMemoryDeviceSizeChange, }; static void -- 2.26.2

This commit adds new memorydevices.rst page which should serve all models of memory devices. Yet, I'm documenting virtio-mem quirks only. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> --- docs/kbase/index.rst | 4 + docs/kbase/memorydevices.rst | 160 +++++++++++++++++++++++++++++++++++ docs/kbase/meson.build | 1 + 3 files changed, 165 insertions(+) create mode 100644 docs/kbase/memorydevices.rst diff --git a/docs/kbase/index.rst b/docs/kbase/index.rst index 532804fe05..45450bf33b 100644 --- a/docs/kbase/index.rst +++ b/docs/kbase/index.rst @@ -46,6 +46,10 @@ Usage `PCI topology <../pci-addresses.html>`__ Addressing schemes for PCI devices +`Memory devices <memorydevices.html>`__ + Memory devices and their use + + Internals / Debugging --------------------- diff --git a/docs/kbase/memorydevices.rst b/docs/kbase/memorydevices.rst new file mode 100644 index 0000000000..23adf54e16 --- /dev/null +++ b/docs/kbase/memorydevices.rst @@ -0,0 +1,160 @@ +============== +Memory devices +============== + +.. contents:: + +Basics +====== + +Memory devices can be divided into two families: DIMMs and NVDIMMs. The former +is typical RAM memory: it's volatile and thus its contents doesn't survive +reboots nor guest shut downs and power ons. The latter retains its contents +across reboots or power outages. + +In Libvirt, there are two models for DIMMs: + +* ``dimm`` model: + + :: + + <memory model='dimm'> + <target> + <size unit='KiB'>523264</size> + <node>0</node> + </target> + <address type='dimm' slot='0'/> + </memory> + +* ``virtio-mem`` model: + + :: + + <memory model='virtio-mem'> + <target> + <size unit='KiB'>1048576</size> + <node>0</node> + <block unit='KiB'>2048</block> + <requested unit='KiB'>524288</requested> + </target> + <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> + </memory> + +Then there are two models for NVDIMMs: + +* ``nvidmm`` model: + + :: + + <memory model='nvdimm'> + <source> + <path>/tmp/nvdimm</path> + </source> + <target> + <size unit='KiB'>523264</size> + <node>0</node> + </target> + <address type='dimm' slot='0'/> + </memory> + +* ``virtio-pmem`` model: + + :: + + <memory model='virtio-pmem' access='shared'> + <source> + <path>/tmp/virtio_pmem</path> + </source> + <target> + <size unit='KiB'>524288</size> + </target> + <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/> + </memory> + + +Please not that (maybe somewhat surprisingly) virtio models go onto PCI bus +instead of DIMM slots. + +Furthermore, DIMMs can have ``<source/>`` element which configures backend for +devices. For NVDIMMs the element is mandatory and reflects where the contents +is saved. + +See https://libvirt.org/formatdomain.html#elementsMemory + +``virtio-mem`` model +==================== + +The ``virtio-mem`` model can be viewed as revised memory balloon. It offers +memory hotplug and hotunplug solution (without the actual hotplug of the +device). It solves problems that memory balloon can't solve on its own and thus +is more flexible than DIMM + balloon solution. ``virtio-mem`` is NUMA aware, +and thus memory can be inflated/deflated only for a subset of guest NUMA nodes. +Also, it works with chunks that are either exposed to guest or taken back from +it. + +See https://virtio-mem.gitlab.io/ + +Under the hood, ``virtio-mem`` device is split into chunks of equal size which +are then exposed to the guest. Either all of them or only a portion depending +on user's request. Therefore there are three important sizes for +``virtio-mem``. All are to be found under ``<target/>`` element: + +#. The maximum size the device can ever offer, exposed under ``<size/>`` +#. The size a single block, exposed under ``<block/>`` +#. The current size exposed to the guest, exposed under ``<requested/>`` + +For instance, the following example the maximum size is 4GiB, the block size is +2MiB and only 1GiB should be exposed to the guest: + + :: + + <memory model='virtio-mem'> + <target> + <size unit='KiB'>4194304</size> + <block unit='KiB'>2048</block> + <requested unit='KiB'>1048576</requested> + </target> + </memory> + +Please note that ``<requested/>`` must be an integer multiple of ``<block/>`` +size or zero (memory completely deflated) and has to be less or equal to +``<size/>`` (memory completely inflated). Furthermore, QEMU recommends the +``<block/>`` size to be as big as a Transparent Huge Page (usually 2MiB). + +To change the size exposed to the guest, users should pass memory device XML +with nothing but ``<requested/>`` changed into the +``virDomainUpdateDeviceFlags()`` API. For user's convenience this can be done +via virsh too: + + :: + + # virsh update-memory-device $dom --requested size 2GiB + +If there are two or more ``<memory/>`` devices then ``--alias`` shall be used +to tell virsh which memory device should be updated. + +For running guests there is fourth size that can be found under ``<target/>``: + + :: + + <actual unit='KiB'>2097152</actual> + +The ``<actual/>`` reflects the actual size consumed by the guest. In general it +can differ from ``<requested/>``. Reasons include guest kernel missing +``virtio-mem`` module and thus being unable to take offered memory, or guest +kernel being unable to free memory and allow deflation. Since ``<actual/>`` +only reports size to users, the element is never parsed. It is formatted only +into live XML. + +Since changing actual allocation requires cooperation with guest kernel, +requests for change are not instant. Therefore, libvirt emits +``VIR_DOMAIN_EVENT_ID_MEMORY_DEVICE_SIZE_CHANGE`` event whenever actual +allocation changed. + +Please not that using ``virtio-mem`` with memory balloon is not possible, +currently. The real reason is that libvirt's memory accounting isn't ready and +mixing these two would be confusing to users. Libvirt exposes current value of +memory balloon under ``<currentMemory/>`` but if it were to account for +``<actual/>`` too then it would be impossible to learn true size of the +balloon. Also it might result in mistakenly trying to deflate ``virtio-mem`` +via ``setmem`` command. diff --git a/docs/kbase/meson.build b/docs/kbase/meson.build index 7b4e7abbd3..7f6adba915 100644 --- a/docs/kbase/meson.build +++ b/docs/kbase/meson.build @@ -9,6 +9,7 @@ docs_kbase_files = [ 'locking-lockd', 'locking', 'locking-sanlock', + 'memorydevices', 'migrationinternals', 'qemu-passthrough-security', 'rpm-deployment', -- 2.26.2