- Devel - libvirt List Archives

[libvirt] [Openstack][Openstack-dev]Where is libvirt library packages in Openstack Nova branch
by harryxiyou 21 Jan '13

21 Jan '13

Hi all, I read the source code of Openstack Nova branch source codes but i can not find the standard libvirt library packages, which i think Nova uses libvirt interfaces they are from standard libvirt library to attach Sheepdog(or others) volumes to QEMU. If i add a new block storage driver for standard libvirt, i have to update these libvirt library packages in Openstack Nova brahch. Cloud anyone give me some suggestions? Thanks in advance ;-) -- Thanks Harry Wei

2 6

[libvirt] [libvirt-java] [PATCH 0/2] fail on test errors and micro optimization
by Claudio Bley 21 Jan '13

21 Jan '13

Hi. This series is pretty self-explanatory, each patch can be applied individually. Tested with JNA 3.3.0, 3.4.0, 3.4.1, 3.4.2, 3.5.0 and 3.5.1. Claudio Bley (2): Make the build fail if any of the JUnit tests have errors or failures Avoid creating new Arrays and Lists on every struct instantiation build.xml | 3 ++- src/main/java/org/libvirt/jna/virConnectAuth.java | 12 +++++++++--- .../java/org/libvirt/jna/virConnectCredential.java | 12 +++++++++--- .../java/org/libvirt/jna/virDomainBlockInfo.java | 12 +++++++++--- .../java/org/libvirt/jna/virDomainBlockStats.java | 12 +++++++++--- src/main/java/org/libvirt/jna/virDomainInfo.java | 12 +++++++++--- .../org/libvirt/jna/virDomainInterfaceStats.java | 12 +++++++++--- .../java/org/libvirt/jna/virDomainJobInfo.java | 12 +++++++++--- .../java/org/libvirt/jna/virDomainMemoryStats.java | 10 ++++++++-- src/main/java/org/libvirt/jna/virError.java | 12 +++++++++--- src/main/java/org/libvirt/jna/virNodeInfo.java | 12 +++++++++--- .../java/org/libvirt/jna/virSchedParameter.java | 12 +++++++++--- .../java/org/libvirt/jna/virStoragePoolInfo.java | 12 +++++++++--- .../java/org/libvirt/jna/virStorageVolInfo.java | 12 +++++++++--- src/main/java/org/libvirt/jna/virVcpuInfo.java | 12 +++++++++--- 15 files changed, 127 insertions(+), 42 deletions(-) -- 1.7.9.5

3 6

[libvirt] [libvirt-java] [PATCH 00/12] event support
by Claudio Bley 21 Jan '13

21 Jan '13

Hi. This patch series is for early review. Please comment. It's not everything set in stone already. For one, names of methods are subject to change. Claudio Bley (12): Introduce event loop support Wrap virConnectIsAlive Wrap virConnectSetKeepAlive Prepare to define proper domain event callback support. Rename method domainEventDeregisterAny to domainEventDeregister Add an internal domainEventRegister forwarding method Add constants for enum virDomainEventID Connect: overload domainEventRegister method for IOError events Connect: overload domainEventRegister method for Reboot events Connect: overload domainEventRegister method for Lifecycle events Connect: overload domainEventRegister for PMWakeup events Connect: overload domainEventRegister method for PMSuspend events src/main/java/org/libvirt/Connect.java | 367 +++++++++++++++++++++-- src/main/java/org/libvirt/Domain.java | 85 ++++++ src/main/java/org/libvirt/jna/Libvirt.java | 53 +++- src/test/java/org/libvirt/TestJavaBindings.java | 1 + 4 files changed, 482 insertions(+), 24 deletions(-) -- 1.7.9.5

1 13

[libvirt] [PATCH v2 00/13] More refactoring of QEMU driver objects & locking
by Daniel P. Berrange 21 Jan '13

21 Jan '13

An update of https://www.redhat.com/archives/libvir-list/2013-January/msg01108.html This series does more general refactoring to help in the goal of killing the big QEMU driver lock. There are 3 key changes in this series - Isolating all QEMU configuration parameters in one struct. - Making virDomainObjList self-locking - Add dedicated locking to PCI/USB device lists There are all the really hard bits of the QEMU locking. As can be seen from the last patch there is still more work to be done before the QEMU driver lock can be said to be truely dead, but this is less critical. In v2: - Rebase, since previous series no longer applied to GIT

1 13

[libvirt] [PATCH 1/1] ppc64 cpu features
by Li Zhang 21 Jan '13

21 Jan '13

From: Li Zhang <zhlcindy(a)linux.vnet.ibm.com> This patch adds a CPU feature "powernv" identifying IBM Power processor that supports native hypervisor e.g. KVM. This can be used by virtualization management software to determine the CPU capabilities. PVR doesn't indicate whether it a host or a guest CPU. So, we use /proc/cpuinfo to get the platform information (PowerNV) and use that to set the "powernv" flag. Signed-off-by: Dipankar Sarma <dipankar(a)in.ibm.com> Signed-off-by: Li Zhang <zhlcindy(a)linux.vnet.ibm.com> --- src/cpu/cpu_map.xml | 9 ++ src/cpu/cpu_powerpc.c | 349 ++++++++++++++++++++++++++++++++++++++---------- src/cpu/cpu_ppc_data.h | 4 + src/util/sysinfo.c | 2 +- 4 files changed, 294 insertions(+), 70 deletions(-) diff --git a/src/cpu/cpu_map.xml b/src/cpu/cpu_map.xml index eb69a34..6b1f9b9 100644 --- a/src/cpu/cpu_map.xml +++ b/src/cpu/cpu_map.xml @@ -587,14 +587,23 @@ <arch name='ppc64'>  <vendor name='IBM' string='PowerPC'/> + <feature name='powernv'>  + <systemid platform='0x00000001'/> + </feature>  <model name='POWER7'> + <feature name='powernv'/> + <systemid pvr='0x003f0200'/> <vendor name='IBM'/> </model> <model name='POWER7_v2.1'> + <feature name='powernv'/> + <systemid pvr='0x003f0201'/> <vendor name='IBM'/> </model> <model name='POWER7_v2.3'> + <feature name='powernv'/> + <systemid pvr='0x003f0203'/> <vendor name='IBM'/> </model> </arch> diff --git a/src/cpu/cpu_powerpc.c b/src/cpu/cpu_powerpc.c index ac10789..2135944 100644 --- a/src/cpu/cpu_powerpc.c +++ b/src/cpu/cpu_powerpc.c @@ -33,6 +33,7 @@ #include "cpu_map.h" #include "buf.h" +#include "sysinfo.h" #define VIR_FROM_THIS VIR_FROM_CPU @@ -44,16 +45,9 @@ struct cpuPowerPC { uint32_t pvr; }; -static const struct cpuPowerPC cpu_defs[] = { - {"POWER7", "IBM", 0x003f0200}, - {"POWER7_v2.1", "IBM", 0x003f0201}, - {"POWER7_v2.3", "IBM", 0x003f0203}, - {NULL, NULL, 0xffffffff} -}; - - struct ppc_vendor { char *name; + uint32_t pvr; struct ppc_vendor *next; }; @@ -64,64 +58,191 @@ struct ppc_model { struct ppc_model *next; }; +struct ppc_feature { + char *name; + union cpuData *data; + struct ppc_feature *next; +}; + struct ppc_map { struct ppc_vendor *vendors; + struct ppc_feature *features; struct ppc_model *models; }; +static void +ppcDataSubtract(union cpuData *data1, + const union cpuData *data2) +{ + data1->ppc.platform &= ~data2->ppc.platform; +} + +static bool +ppcDataIsSubset(const union cpuData *data, + const union cpuData *subset) +{ + if (subset->ppc.platform && + (data->ppc.platform & subset->ppc.platform) == subset->ppc.platform) + return true; + + return false; +} + +static void +PowerPCDataFree(union cpuData *data) +{ + if (data == NULL) + return; + VIR_FREE(data); +} + + +static union cpuData * +ppcDataCopy(const union cpuData *data) +{ + union cpuData *copy = NULL; + + if (VIR_ALLOC(copy) < 0) { + PowerPCDataFree(copy); + return NULL; + } + copy->ppc = data->ppc; + return copy; +} + + +/* also removes all detected features from data */ static int -ConvertModelVendorFromPVR(char ***model, - char ***vendor, - uint32_t pvr) +ppcDataToCPUFeatures(virCPUDefPtr cpu, + int policy, + union cpuData *data, + const struct ppc_map *map) { - int i; + const struct ppc_feature *feature = map->features; - for (i = 0; cpu_defs[i].name; i++) { - if (cpu_defs[i].pvr == pvr) { - **model = strdup(cpu_defs[i].name); - **vendor = strdup(cpu_defs[i].vendor); - return 0; + while (feature != NULL) { + if (ppcDataIsSubset(data, feature->data)) { + ppcDataSubtract(data, feature->data); + if (virCPUDefAddFeature(cpu, feature->name, policy) < 0) + return -1; } + feature = feature->next; } - virReportError(VIR_ERR_INTERNAL_ERROR, - "%s", _("Missing the definition of this model")); - return -1; + return 0; } -static int -ConvertPVRFromModel(const char *model, - uint32_t *pvr) +static struct ppc_feature * +ppcFeatureNew(void) { - int i; + struct ppc_feature *feature; - for (i = 0; cpu_defs[i].name; i++) { - if (STREQ(cpu_defs[i].name, model)) { - *pvr = cpu_defs[i].pvr; - return 0; - } + if (VIR_ALLOC(feature) < 0) + return NULL; + + if (VIR_ALLOC(feature->data) < 0) { + VIR_FREE(feature); + return NULL; + } + + return feature; +} + + +static void +ppcFeatureFree(struct ppc_feature *feature) +{ + if (feature == NULL) + return; + + VIR_FREE(feature->name); + PowerPCDataFree(feature->data); + VIR_FREE(feature); +} + + +static struct ppc_feature * +ppcFeatureFind(const struct ppc_map *map, + const char *name) +{ + struct ppc_feature *feature; + + feature = map->features; + while (feature != NULL) { + if (STREQ(feature->name, name)) + return feature; + + feature = feature->next; } - virReportError(VIR_ERR_INTERNAL_ERROR, - "%s", _("Missing the definition of this model")); - return -1; + return NULL; } static int -cpuMatch(const union cpuData *data, - char **cpu_model, - char **cpu_vendor, - const struct ppc_model *model) +ppcFeatureLoad(xmlXPathContextPtr ctxt, + struct ppc_map *map) { + xmlNodePtr *nodes = NULL; + xmlNodePtr ctxt_node = ctxt->node; + struct ppc_feature *feature; int ret = 0; + unsigned long platform = 0; + unsigned long pvr = 0; + int ret_platform; + int ret_pvr; + int n; + + if (!(feature = ppcFeatureNew())) + goto no_memory; + + feature->name = virXPathString("string(@name)", ctxt); + if (feature->name == NULL) { + virReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("Missing CPU feature name")); + goto ignore; + } + + if (ppcFeatureFind(map, feature->name)) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("CPU feature %s already defined"), feature->name); + goto ignore; + } + + n = virXPathNodeSet("./systemid", ctxt, &nodes); + if (n < 0) + goto ignore; + + ctxt->node = nodes[0]; + ret_platform = virXPathULongHex("string(@platform)", ctxt, &platform); + ret_pvr = virXPathULongHex("string(@pvr)", ctxt, &pvr); + if (ret_platform < 0 && ret_pvr < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Invalid systemid in %s feature"), feature->name); + goto ignore; + } + feature->data->ppc.platform = platform; + feature->data->ppc.pvr = pvr; - ret = ConvertModelVendorFromPVR(&cpu_model, &cpu_vendor, data->ppc.pvr); + if (map->features == NULL) + map->features = feature; + else { + feature->next = map->features; + map->features = feature; + } - if (STREQ(model->name, *cpu_model) && - STREQ(model->vendor->name, *cpu_vendor)) - ret = 1; +out: + ctxt->node = ctxt_node; + VIR_FREE(nodes); return ret; + +no_memory: + virReportOOMError(); + ret = -1; + +ignore: + ppcFeatureFree(feature); + goto out; } @@ -171,6 +292,66 @@ ppcModelFind(const struct ppc_map *map, return NULL; } +/* also removes bits corresponding to vendor string from data */ +static const struct ppc_vendor * +ppcDataToVendor(union cpuData *data, + const struct ppc_map *map) +{ + const struct ppc_vendor *vendor = map->vendors; + + while (vendor) { + if (data->ppc.pvr == vendor->pvr) + return vendor; + vendor = vendor->next; + } + + return NULL; +} + + +static virCPUDefPtr +ppcDataToCPU(const union cpuData *data, + const struct ppc_model *model, + const struct ppc_map *map) +{ + virCPUDefPtr cpu; + union cpuData *copy = NULL; + union cpuData *modelData = NULL; + const struct ppc_vendor *vendor; + + if (VIR_ALLOC(cpu) < 0 || + !(cpu->model = strdup(model->name)) || + !(copy = ppcDataCopy(data)) || + !(modelData = ppcDataCopy(model->data))) + goto no_memory; + + if ((vendor = ppcDataToVendor(copy, map)) && + !(cpu->vendor = strdup(vendor->name))) + goto no_memory; + + ppcDataSubtract(copy, modelData); + ppcDataSubtract(modelData, data); + + /* because feature policy is ignored for host CPU */ + cpu->type = VIR_CPU_TYPE_GUEST; + + if (ppcDataToCPUFeatures(cpu, VIR_CPU_FEATURE_REQUIRE, copy, map) || + ppcDataToCPUFeatures(cpu, VIR_CPU_FEATURE_DISABLE, modelData, map)) + goto error; + +cleanup: + PowerPCDataFree(modelData); + PowerPCDataFree(copy); + return cpu; + +no_memory: + virReportOOMError(); +error: + virCPUDefFree(cpu); + cpu = NULL; + goto cleanup; +} + static struct ppc_vendor * ppcVendorFind(const struct ppc_map *map, const char *name) @@ -256,6 +437,9 @@ ppcModelLoad(xmlXPathContextPtr ctxt, xmlNodePtr *nodes = NULL; struct ppc_model *model; char *vendor = NULL; + unsigned long pvr = 0, platform = 0; + int ret_platform, ret_pvr; + int n; int ret = -1; if (!(model = ppcModelNew())) @@ -268,10 +452,20 @@ ppcModelLoad(xmlXPathContextPtr ctxt, goto ignore; } - ret = ConvertPVRFromModel(model->name, &model->data->ppc.pvr); - if (ret < 0) - goto ignore; + n = virXPathNodeSet("./systemid", ctxt, &nodes); + if (n < 0) + goto ignore; + ctxt->node = nodes[0]; + ret_platform = virXPathULongHex("string(@platform)", ctxt, &platform); + ret_pvr = virXPathULongHex("string(@pvr)", ctxt, &pvr); + if (ret_platform < 0 && ret_pvr < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Invalid systemid in %s model"), model->name); + goto ignore; + } + model->data->ppc.platform = platform; + model->data->ppc.pvr = pvr; if (virXPathBoolean("boolean(./vendor)", ctxt)) { vendor = virXPathString("string(./vendor/@name)", ctxt); @@ -324,6 +518,8 @@ ppcMapLoadCallback(enum cpuMapElement element, return ppcVendorLoad(ctxt, map); case CPU_MAP_ELEMENT_MODEL: return ppcModelLoad(ctxt, map); + case CPU_MAP_ELEMENT_FEATURE: + return ppcFeatureLoad(ctxt, map); default: break; } @@ -385,6 +581,7 @@ ppcModelCopy(const struct ppc_model *model) } copy->data->ppc.pvr = model->data->ppc.pvr; + copy->data->ppc.platform = model->data->ppc.platform; copy->vendor = model->vendor; return copy; @@ -437,8 +634,6 @@ PowerPCDecode(virCPUDefPtr cpu, const struct ppc_model *candidate; virCPUDefPtr cpuCandidate; virCPUDefPtr cpuModel = NULL; - char *cpu_vendor = NULL; - char *cpu_model = NULL; unsigned int i; if (data == NULL || (map = ppcLoadMap()) == NULL) @@ -475,13 +670,30 @@ PowerPCDecode(virCPUDefPtr cpu, goto next; } - if (VIR_ALLOC(cpuCandidate) < 0) { - virReportOOMError(); + if (!(cpuCandidate = ppcDataToCPU(data, candidate, map))) goto out; + + if (candidate->vendor && cpuCandidate->vendor && + STRNEQ(candidate->vendor->name, cpuCandidate->vendor)) { + VIR_DEBUG("CPU vendor %s of model %s differs from %s; ignoring", + candidate->vendor->name, candidate->name, + cpuCandidate->vendor); + virCPUDefFree(cpuCandidate); + goto next; } - cpuCandidate->model = strdup(candidate->name); - cpuCandidate->vendor = strdup(candidate->vendor->name); + if (cpu->type == VIR_CPU_TYPE_HOST) { + cpuCandidate->type = VIR_CPU_TYPE_HOST; + for (i = 0; i < cpuCandidate->nfeatures; i++) { + switch (cpuCandidate->features[i].policy) { + case VIR_CPU_FEATURE_DISABLE: + virCPUDefFree(cpuCandidate); + goto next; + default: + cpuCandidate->features[i].policy = -1; + } + } + } if (preferred && STREQ(cpuCandidate->model, preferred)) { virCPUDefFree(cpuModel); @@ -489,19 +701,12 @@ PowerPCDecode(virCPUDefPtr cpu, break; } - ret = cpuMatch(data, &cpu_model, &cpu_vendor, candidate); - if (ret < 0) { - VIR_FREE(cpuCandidate); - goto out; - } else if (ret == 1) { - cpuCandidate->model = cpu_model; - cpuCandidate->vendor = cpu_vendor; + if (cpuModel == NULL + || cpuModel->nfeatures > cpuCandidate->nfeatures) { virCPUDefFree(cpuModel); cpuModel = cpuCandidate; - break; - } - - virCPUDefFree(cpuCandidate); + }else + virCPUDefFree(cpuCandidate); next: candidate = candidate->next; @@ -515,6 +720,8 @@ PowerPCDecode(virCPUDefPtr cpu, cpu->model = cpuModel->model; cpu->vendor = cpuModel->vendor; + cpu->nfeatures = cpuModel->nfeatures; + cpu->features = cpuModel->features; VIR_FREE(cpuModel); ret = 0; @@ -537,19 +744,11 @@ static uint32_t ppc_mfpvr(void) } #endif -static void -PowerPCDataFree(union cpuData *data) -{ - if (data == NULL) - return; - - VIR_FREE(data); -} - static union cpuData * PowerPCNodeData(void) { union cpuData *data; + virSysinfoDefPtr hostinfo; if (VIR_ALLOC(data) < 0) { virReportOOMError(); @@ -561,6 +760,18 @@ PowerPCNodeData(void) data->ppc.pvr = ppc_mfpvr(); #endif + hostinfo = virSysinfoRead(); + if (hostinfo == NULL) { + VIR_FREE(data); + return NULL; + } + + data->ppc.platform &= ~VIR_CPU_PPC64_POWERNV; + + if (STREQ(hostinfo->system_family, "PowerNV")) + data->ppc.platform |= VIR_CPU_PPC64_POWERNV; + virSysinfoDefFree(hostinfo); + return data; } diff --git a/src/cpu/cpu_ppc_data.h b/src/cpu/cpu_ppc_data.h index 685332a..0e691ce 100644 --- a/src/cpu/cpu_ppc_data.h +++ b/src/cpu/cpu_ppc_data.h @@ -28,6 +28,10 @@ struct cpuPPCData { uint32_t pvr; + uint32_t platform; /* Bitflag indicating platform features */ }; +#define VIR_CPU_PPC64_NONE 0x0 +#define VIR_CPU_PPC64_POWERNV 0x1 + #endif /* __VIR_CPU_PPC_DATA_H__ */ diff --git a/src/util/sysinfo.c b/src/util/sysinfo.c index 6a5db80..5f98986 100644 --- a/src/util/sysinfo.c +++ b/src/util/sysinfo.c @@ -234,7 +234,7 @@ virSysinfoRead(void) { if (VIR_ALLOC(ret) < 0) goto no_memory; - if (virFileReadAll(CPUINFO, 2048, &outbuf) < 0) { + if (virFileReadAll(CPUINFO, 8192, &outbuf) < 0) { virReportError(VIR_ERR_INTERNAL_ERROR, _("Failed to open %s"), CPUINFO); return NULL; -- 1.7.10.1

1 1

[libvirt] Network stops with CentOS 6.3 guest on Ubuntu 12.10
by Andy Howell 20 Jan '13

20 Jan '13

Hello, I'm trying to get CentOS 6.3 running under Ubuntu 12.10. When the interface comes up, I can ping another machine for 9 packets. After that it stops working. Bouncing the guest network interface gets it going again for another 10 packets. Monitoring the vnet0 interface on the host, I see ARP requests from the host to the guest are not getting replies. I cleared out iptables to eliminate that. No luck. I have Fedora 17 and Windows 7 running on the machine with no problems. Any ideas? Thanks, Andy

1 0

[libvirt] Fwd: [Openstack] [Openstack-dev][Sheepdog]Add a new driver for OpenstackCinder like Sheepdog volumes
by harryxiyou 19 Jan '13

19 Jan '13

---------- Forwarded message ---------- From: harryxiyou <harryxiyou(a)gmail.com> Date: Sat, Jan 19, 2013 at 11:49 AM Subject: Re: [Openstack] [Openstack-dev][Sheepdog]Add a new driver for OpenstackCinder like Sheepdog volumes To: 张家龙 <zhangjl(a)awcloud.com> Cc: Kazutaka Morita <morita.kazutaka(a)gmail.com>, openstack-dev(a)lists.launchpad.org, Liu Yuan <namei.unix(a)gmail.com>, sheepdog <sheepdog(a)lists.wpkg.org>, openstack <openstack(a)lists.launchpad.net>, Kang Hua <kanghua151(a)gmail.com> On Sat, Jan 19, 2013 at 10:57 AM, 张家龙 <zhangjl(a)awcloud.com> wrote: > Maybe，you need to modify the nova/virt/libvirt.xml.template.In our tesing of > nova and sheepdog, Hi Zhang jia long, What are the functions of libvirt.xml.template? Currently, we just wanna to add a new block driver(a new one as the same as sheepdog driver) to Openstack Cinder ;-) The performances will be enhenced later. > [...] > Hi Morita and other developers, > > If i add a QEMU/Libvirt driver(the same as Sheepdog volumes driver in > Openstack Cinder branch) to let Openstack Cinder support a new block-level > storage system, I should change following stuffs, right? > > 1, Add a driver file to the dir in Openstack Cinder branch(the same as > sheepdog.py), > https://github.com/openstack/cinder/blob/master/cinder/volume/drivers > 2, Change the file in Openstack Nova branch(Let libvirt attach HLFS volumes > to QEMU, the same as sheepdog)， > > https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py > > Do I should change/add any other files for Openstack? Cloud anyone give me > some suggestions? Thanks in advance ;-) > Hi Openstack developers, libvirt developers and Sheepdog developers, I also have following questions, please give me some suggestions. Thanks in advance ;-) 1, The libvirt codes in Nova branch are just interfaces which call the standard Libvirt library, right? (Or the libvirt codes in Nova branch are the standard Libvirt library.) 2, If i add a new block storage driver for Openstack by QEMU/Libvirt way, should i also modify standard libvirt library and QEMU source codes to let them support this new driver so that i can use this new driver to let libvirt attach the new driver volumes to QEMU in Openstack? 3, Certainly, i should also do following stuffs. a, Add a driver file to the dir in Openstack Cinder branch(the same as sheepdog.py), https://github.com/openstack/cinder/blob/master/cinder/volume/drivers b, Change the file in Openstack Nova branch(Let libvirt attach the new driver volumes to QEMU, the same as sheepdog)， https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py 4, Should i have to do any other stuffs? -- Thanks Harry Wei -- Thanks Harry Wei

1 0

[libvirt] [PATCH] Fix race condition when destroying guests
by Daniel P. Berrange 19 Jan '13

19 Jan '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> When running virDomainDestroy, we need to make sure that no other background thread cleans up the domain while we're doing our work. This can happen if we release the domain object while in the middle of work, because the monitor might detect EOF in this window. For this reason we have a 'beingDestroyed' flag to stop the monitor from doing its normal cleanup. Unfortunately this flag was only being used to protect qemuDomainBeginJob, and not qemuProcessKill This left open a race condition where either libvirtd could crash, or alternatively report bogus error messages about the domain already having been destroyed to the caller Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/qemu/qemu_driver.c | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index c28c223..2a6f381 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2121,24 +2121,29 @@ qemuDomainDestroyFlags(virDomainPtr dom, qemuDomainSetFakeReboot(driver, vm, false); + + /* We need to prevent monitor EOF callback from doing our work (and sending + * misleading events) while the vm is unlocked inside BeginJob/ProcessKill API + */ + priv->beingDestroyed = true; + /* Although qemuProcessStop does this already, there may * be an outstanding job active. We want to make sure we * can kill the process even if a job is active. Killing * it now means the job will be released */ if (flags & VIR_DOMAIN_DESTROY_GRACEFUL) { - if (qemuProcessKill(driver, vm, 0) < 0) + if (qemuProcessKill(driver, vm, 0) < 0) { + priv->beingDestroyed = false; goto cleanup; + } } else { - if (qemuProcessKill(driver, vm, VIR_QEMU_PROCESS_KILL_FORCE) < 0) + if (qemuProcessKill(driver, vm, VIR_QEMU_PROCESS_KILL_FORCE) < 0) { + priv->beingDestroyed = false; goto cleanup; + } } - /* We need to prevent monitor EOF callback from doing our work (and sending - * misleading events) while the vm is unlocked inside BeginJob API - */ - priv->beingDestroyed = true; - if (qemuDomainObjBeginJobWithDriver(driver, vm, QEMU_JOB_DESTROY) < 0) goto cleanup; -- 1.8.0.2

5 6

[libvirt] RFC: libxl race fixes
by Jim Fehlig 18 Jan '13

18 Jan '13

I've been investigating some races in the libxl driver and would like to get comments on some potential solutions. The first race is in the fd/timeout event handling code, which maps libxl's osevent interface to libvirt's event loop interface. This mapping opens the possibility for libvirt's event loop to invoke event callbacks after libxl has already deregistered the event, potentially accessing an event object that has already been freed. One solution to this race I've found successful is reference counting the objects associated with the events. When libxl registers an event, an object encapsulating the event is created and it's reference count is set to 1. When the event is injected into libvirt's event loop, another reference is taken on the object. When libxl deregisters the event, it's reference count is decremented. Once the event is removed from libvirt's event loop, the final reference is decremented and the object is disposed. This approach ensures the object is not disposed until it is removed from libvirt's event loop *and* libxl had explicitly deregistered the event. The notion of an event being 'disabled' found in libvirt's event loop impl also had to be added to the libxl event object, to ensure the driver doesn't call into libxl for a previously deregistered event. The second race is between destroying a vm (i.e., calling privateDataFreeFunc, which frees the libxl ctx) and deregistration/cleanup of all events that have been registered by libxl. One solution for this race is to convert libxlDomainObjPrivate to a virObject and increment its reference for each fd and timeout registration. Only when all fds and timeouts are deregistered and destroyed will the libxlDomainObjPrivate be destroyed. One downside to this approach is that an API to "cleanup" the libxl ctx is needed. Without such an API, some fds are not deregistered until calling libxl_ctx_free. But since the fd events have references on the libxlDomainObjPrivate, libxl_ctx_free is never called. (BTW I have a patch adding libxl_ctx_quiesce() to libxl, which upstream xen folks seem receptive to, including backporting to Xen 4.2 branch, but I don't think this is ideal.) An alternate solution that can be used to address both of these races is to maintain a list of the fd/timeout registrations in the libxlDomainObjPrivate object, and take a more brute force approach to managing the registrations Other solutions that I might be overlooking are certainly appreciated :). Regards, Jim

2 3

[libvirt] [PATCH] qemu: Support ram bar size for qxl devices
by Alon Levy 18 Jan '13

18 Jan '13

Adds a qxl-ram attribute globaly to the video.model element, that changes the resulting qemu command line only if video.type == "qxl". That attribute gets a default value of 64*1024 only if model.type is "qxl". In effect not changing any xml or argv for non qxl devices. For qxl devices a new property is set: -global qxl-vga.ram_size=<ram>*1024 or -global qxl.ram_size=<ram>*1024 For the main and secondary qxl devices respectively. The default for the qxl ram bar is the same as the default for the qxl vram bar, 64*1024. --- I've added a qxl-ram attribute. There is no precedent for adding am attribute prefixed like this, so I'm open for any other suggestion on how to do it. docs/schemas/domaincommon.rng | 9 +++- src/conf/domain_conf.c | 19 ++++++- src/conf/domain_conf.h | 1 + src/qemu/qemu_command.c | 58 ++++++++++++++++++---- .../qemuxml2argv-graphics-spice-compression.args | 2 +- .../qemuxml2argv-graphics-spice-compression.xml | 4 +- .../qemuxml2argv-graphics-spice-qxl-vga.args | 2 +- .../qemuxml2argv-graphics-spice-qxl-vga.xml | 4 +- .../qemuxml2argv-graphics-spice.args | 2 +- .../qemuxml2argv-graphics-spice.xml | 4 +- .../qemuxml2argv-video-device-pciaddr-default.args | 6 +-- 11 files changed, 86 insertions(+), 25 deletions(-) diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index 67ae864..50fc834 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -2251,7 +2251,9 @@ </define>  <define name="video"> <element name="video"> @@ -2268,6 +2270,11 @@ </choice> </attribute> <optional> + <attribute name="qxl-ram"> + <ref name="unsignedInt"/> + </attribute> + </optional> + <optional> <attribute name="vram"> <ref name="unsignedInt"/> </attribute> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 13c14e9..a16f1b5 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -7391,6 +7391,7 @@ virDomainVideoDefParseXML(const xmlNodePtr node, char *type = NULL; char *heads = NULL; char *vram = NULL; + char *qxl_ram = NULL; char *primary = NULL; if (VIR_ALLOC(def) < 0) { @@ -7401,9 +7402,10 @@ virDomainVideoDefParseXML(const xmlNodePtr node, cur = node->children; while (cur != NULL) { if (cur->type == XML_ELEMENT_NODE) { - if (!type && !vram && !heads && + if (!type && !vram && !qxl_ram && !heads && xmlStrEqual(cur->name, BAD_CAST "model")) { type = virXMLPropString(cur, "type"); + qxl_ram = virXMLPropString(cur, "qxl-ram"); vram = virXMLPropString(cur, "vram"); heads = virXMLPropString(cur, "heads"); @@ -7431,6 +7433,18 @@ virDomainVideoDefParseXML(const xmlNodePtr node, } } + if (qxl_ram) { + if (virStrToLong_ui(qxl_ram, NULL, 10, &def->qxl_ram) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("cannot parse video qxl-ram '%s'"), qxl_ram); + goto error; + } + } else { + if (def->type == VIR_DOMAIN_VIDEO_TYPE_QXL) { + def->qxl_ram = virDomainVideoDefaultRAM(dom, def->type); + } + } + if (vram) { if (virStrToLong_ui(vram, NULL, 10, &def->vram) < 0) { virReportError(VIR_ERR_INTERNAL_ERROR, @@ -7455,6 +7469,7 @@ virDomainVideoDefParseXML(const xmlNodePtr node, goto error; VIR_FREE(type); + VIR_FREE(qxl_ram); VIR_FREE(vram); VIR_FREE(heads); @@ -13383,6 +13398,8 @@ virDomainVideoDefFormat(virBufferPtr buf, virBufferAddLit(buf, " <video>\n"); virBufferAsprintf(buf, " <model type='%s'", model); + if (def->qxl_ram && def->type == VIR_DOMAIN_VIDEO_TYPE_QXL) + virBufferAsprintf(buf, " qxl-ram='%u'", def->qxl_ram); if (def->vram) virBufferAsprintf(buf, " vram='%u'", def->vram); if (def->heads) diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index ce36003..9333d26 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -1157,6 +1157,7 @@ struct _virDomainVideoAccelDef { struct _virDomainVideoDef { int type; + unsigned int qxl_ram; unsigned int vram; unsigned int heads; bool primary; diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 981c692..05f656a 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -3563,6 +3563,15 @@ qemuBuildDeviceVideoStr(virDomainVideoDefPtr video, UINT_MAX / 1024); goto error; } + if (video->qxl_ram > (UINT_MAX / 1024)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("value for 'qxl-ram' must be less than '%u'"), + UINT_MAX / 1024); + goto error; + } + + /* QEMU accepts bytes for ram_size. */ + virBufferAsprintf(&buf, ",ram_size=%u", video->qxl_ram * 1024); /* QEMU accepts bytes for vram_size. */ virBufferAsprintf(&buf, ",vram_size=%u", video->vram * 1024); @@ -6569,23 +6578,48 @@ qemuBuildCommandLine(virConnectPtr conn, virCommandAddArgList(cmd, "-vga", vgastr, NULL); if (def->videos[0]->type == VIR_DOMAIN_VIDEO_TYPE_QXL) { - if (def->videos[0]->vram && + if ((def->videos[0]->vram || def->videos[0]->qxl_ram) && qemuCapsGet(caps, QEMU_CAPS_DEVICE)) { - if (def->videos[0]->vram > (UINT_MAX / 1024)) { + int qxl_ram = def->videos[0]->qxl_ram; + int vram = def->videos[0]->vram; + if (vram > (UINT_MAX / 1024)) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("value for 'vram' must be less than '%u'"), + _("value for 'vram' must be less than '%u'"), + UINT_MAX / 1024); + goto error; + } + if (qxl_ram > (UINT_MAX / 1024)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("value for 'qxl-ram' must be less than '%u'"), UINT_MAX / 1024); goto error; } - virCommandAddArg(cmd, "-global"); - - if (qemuCapsGet(caps, QEMU_CAPS_DEVICE_QXL_VGA)) - virCommandAddArgFormat(cmd, "qxl-vga.vram_size=%u", - def->videos[0]->vram * 1024); - else - virCommandAddArgFormat(cmd, "qxl.vram_size=%u", - def->videos[0]->vram * 1024); + if (qemuCapsGet(caps, QEMU_CAPS_DEVICE_QXL_VGA)) { + if (qxl_ram) { + virCommandAddArg(cmd, "-global"); + virCommandAddArgFormat(cmd, + "qxl-vga.ram_size=%u", + qxl_ram * 1024); + } + if (vram) { + virCommandAddArg(cmd, "-global"); + virCommandAddArgFormat(cmd, + "qxl-vga.vram_size=%u", + vram * 1024); + } + } else { + if (qxl_ram) { + virCommandAddArg(cmd, "-global"); + virCommandAddArgFormat(cmd, "qxl.ram_size=%u", + qxl_ram * 1024); + } + if (vram) { + virCommandAddArg(cmd, "-global"); + virCommandAddArgFormat(cmd, "qxl.vram_size=%u", + vram * 1024); + } + } } } } @@ -9247,6 +9281,8 @@ virDomainDefPtr qemuParseCommandLine(virCapsPtr caps, else vid->type = video; vid->vram = virDomainVideoDefaultRAM(def, vid->type); + vid->qxl_ram = vid->type == VIR_DOMAIN_VIDEO_TYPE_QXL ? + virDomainVideoDefaultRAM(def, vid->type) : 0; vid->heads = 1; if (VIR_REALLOC_N(def->videos, def->nvideos+1) < 0) { diff --git a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-compression.args b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-compression.args index 5c5912b..2f537df 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-compression.args +++ b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-compression.args @@ -5,5 +5,5 @@ unix:/tmp/test-monitor,server,nowait -no-acpi -boot c -usb -hda \ x509-dir=/etc/pki/libvirt-spice,\ image-compression=auto_glz,jpeg-wan-compression=auto,zlib-glz-wan-compression=auto,\ playback-compression=on,streaming-video=filter -vga \ -qxl -global qxl.vram_size=18874368 -device qxl,id=video1,vram_size=33554432,bus=pci.0,addr=0x4 \ +qxl -global qxl.ram_size=67108864 -global qxl.vram_size=18874368 -device qxl,id=video1,ram_size=67108864,vram_size=33554432,bus=pci.0,addr=0x4 \ -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 diff --git a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-compression.xml b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-compression.xml index 52eb5b9..32e4cc8 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-compression.xml +++ b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-compression.xml @@ -31,10 +31,10 @@ <streaming mode='filter'/> </graphics> <video> - <model type='qxl' vram='18432' heads='1'/> + <model type='qxl' qxl-ram='65536' vram='18432' heads='1'/> </video> <video> - <model type='qxl' vram='32768' heads='1'/> + <model type='qxl' qxl-ram='65536' vram='32768' heads='1'/> </video> <memballoon model='virtio'/> </devices> diff --git a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-qxl-vga.args b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-qxl-vga.args index 3954c03..007dd3f 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-qxl-vga.args +++ b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-qxl-vga.args @@ -3,5 +3,5 @@ LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test QEMU_AUDIO_DRV=spice \ unix:/tmp/test-monitor,server,nowait -no-acpi -boot c -usb -hda \ /dev/HostVG/QEMUGuest1 -spice port=5903,tls-port=5904,addr=127.0.0.1,\ x509-dir=/etc/pki/libvirt-spice,tls-channel=main,plaintext-channel=inputs -vga \ -qxl -global qxl-vga.vram_size=33554432 -device qxl,id=video1,vram_size=67108864,bus=pci.0,addr=0x4 \ +qxl -global qxl-vga.ram_size=67108864 -global qxl-vga.vram_size=33554432 -device qxl,id=video1,ram_size=67108864,vram_size=67108864,bus=pci.0,addr=0x4 \ -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 diff --git a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-qxl-vga.xml b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-qxl-vga.xml index 49cb8cc..ab40c7b 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-qxl-vga.xml +++ b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice-qxl-vga.xml @@ -28,10 +28,10 @@ <channel name='inputs' mode='insecure'/> </graphics> <video> - <model type='qxl' vram='32768' heads='1'/> + <model type='qxl' qxl-ram='65536' vram='32768' heads='1'/> </video> <video> - <model type='qxl' vram='65536' heads='1'/> + <model type='qxl' qxl-ram='65536' vram='65536' heads='1'/> </video> <memballoon model='virtio'/> </devices> diff --git a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice.args b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice.args index 854e723..030e1f2 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice.args +++ b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice.args @@ -5,5 +5,5 @@ unix:/tmp/test-monitor,server,nowait -no-acpi -boot c -usb -hda \ x509-dir=/etc/pki/libvirt-spice,tls-channel=default,tls-channel=main,plaintext-channel=inputs,\ image-compression=auto_glz,jpeg-wan-compression=auto,zlib-glz-wan-compression=auto,\ playback-compression=on,streaming-video=filter,disable-copy-paste -vga \ -qxl -global qxl.vram_size=18874368 -device qxl,id=video1,vram_size=33554432,bus=pci.0,addr=0x4 \ +qxl -global qxl.ram_size=67108864 -global qxl.vram_size=18874368 -device qxl,id=video1,ram_size=67108864,vram_size=33554432,bus=pci.0,addr=0x4 \ -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 diff --git a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice.xml b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice.xml index d4939a4..b540915 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice.xml +++ b/tests/qemuxml2argvdata/qemuxml2argv-graphics-spice.xml @@ -34,10 +34,10 @@ <clipboard copypaste='no'/> </graphics> <video> - <model type='qxl' vram='18432' heads='1'/> + <model type='qxl' qxl-ram='65536' vram='18432' heads='1'/> </video> <video> - <model type='qxl' vram='32768' heads='1'/> + <model type='qxl' qxl-ram='65536' vram='32768' heads='1'/> </video> <memballoon model='virtio'/> </devices> diff --git a/tests/qemuxml2argvdata/qemuxml2argv-video-device-pciaddr-default.args b/tests/qemuxml2argvdata/qemuxml2argv-video-device-pciaddr-default.args index 9ce852f..4abd7c2 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-video-device-pciaddr-default.args +++ b/tests/qemuxml2argvdata/qemuxml2argv-video-device-pciaddr-default.args @@ -3,7 +3,7 @@ LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test QEMU_AUDIO_DRV=none \ -monitor unix:/tmp/test-monitor,server,nowait -no-acpi -boot c \ -usb \ -hda /var/lib/libvirt/images/QEMUGuest1 -vnc 127.0.0.1:-5900 \ --device qxl-vga,id=video0,vram_size=67108864,bus=pci.0,addr=0x3 \ --device qxl,id=video1,vram_size=67108864,bus=pci.0,addr=0x4 \ --device qxl,id=video2,vram_size=67108864,bus=pci.0,addr=0x5 \ +-device qxl-vga,id=video0,ram_size=67108864,vram_size=67108864,bus=pci.0,addr=0x3 \ +-device qxl,id=video1,ram_size=67108864,vram_size=67108864,bus=pci.0,addr=0x4 \ +-device qxl,id=video2,ram_size=67108864,vram_size=67108864,bus=pci.0,addr=0x5 \ -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x2 -- 1.8.0.1

6 10