March 2012 - Devel - Libvirt List Archives

[libvirt] [PATCH 0/5 0/1 0/1 V3] Add new public API virDomainGetPcpusUsage() and pcpuinfo command in virsh

by Lai Jiangshan

"virt-top -1" can call virDomainGetPcpusUsage() periodically and get the CPU activities per CPU. (See the last patch in this series). virsh is also added a pcpuinfo command which calls virDomainGetPcpusUsage(), it gets information about the physical CPUs, such as the usage of CPUs, the current attached vCPUs. # virsh pcpuinfo rhel6 CPU: 0 Curr VCPU: - Usage: 47.3 CPU: 1 Curr VCPU: 1 Usage: 46.8 CPU: 2 Curr VCPU: 0 Usage: 52.7 CPU: 3 Curr VCPU: - Usage: 44.1 Changed from V2: Simple cleanup Add python implementation of virDomainGetPcpusUsage() Acked-by: "Richard W.M. Jones" <rjones(a)redhat.com> Signed-off-by: Lai Jiangshan <laijs(a)cn.fujitsu.com> Patch for libvirt(5 patches): daemon/remote.c | 68 ++++++++++++++++++++++++++++ include/libvirt/libvirt.h.in | 5 ++ python/generator.py | 1 + python/libvirt-override-api.xml | 6 +++ python/libvirt-override.c | 33 ++++++++++++++ src/driver.h | 7 +++ src/libvirt.c | 51 +++++++++++++++++++++ src/libvirt_public.syms | 5 ++ src/qemu/qemu.conf | 5 +- src/qemu/qemu_conf.c | 3 +- src/qemu/qemu_driver.c | 74 +++++++++++++++++++++++++++++++ src/remote/remote_driver.c | 51 +++++++++++++++++++++ src/remote/remote_protocol.x | 17 +++++++- src/remote_protocol-structs | 13 +++++ src/util/cgroup.c | 7 +++ src/util/cgroup.h | 1 + tools/virsh.c | 93 +++++++++++++++++++++++++++++++++++++++ tools/virsh.pod | 5 ++ 18 files changed, 441 insertions(+), 4 deletions(-) Patch for ocaml-libvirt (1 patch): libvirt/libvirt.ml | 1 + libvirt/libvirt.mli | 4 ++++ libvirt/libvirt_c_oneoffs.c | 25 +++++++++++++++++++++++++ 3 files changed, 30 insertions(+), 0 deletions(-) Patch for virt-top (1 patch): virt-top/virt_top.ml | 75 +++++++++++++++++-------------------------------- 1 files changed, 26 insertions(+), 49 deletions(-) -- 1.7.4.4

12 years, 9 months

7
60
0 / 0

[libvirt] [PATCH] util: fix build mingw (and all non-linux) build failure

by Laine Stump

ATTRIBUTE_UNUSED was accidentally forgotten on one arg of a stub function for functionality that's not present on non-linux platforms. This causes a non-linux build with --enable-compile-warnings=error to fail. --- Pushed under build-breaker rule. src/util/pci.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/util/pci.c b/src/util/pci.c index c8a5287..a6212f2 100644 --- a/src/util/pci.c +++ b/src/util/pci.c @@ -2219,7 +2219,8 @@ pciDeviceNetName(char *device_link_sysfs_path ATTRIBUTE_UNUSED, int pciDeviceGetVirtualFunctionInfo(const char *vf_sysfs_device_path ATTRIBUTE_UNUSED, - char **pfname, int *vf_index ATTRIBUTE_UNUSED) + char **pfname ATTRIBUTE_UNUSED, + int *vf_index ATTRIBUTE_UNUSED) { pciReportError(VIR_ERR_INTERNAL_ERROR, _("pciDeviceGetVirtualFunctionInfo " "is not supported on non-linux platforms")); -- 1.7.7.6

12 years, 9 months

1
0
0 / 0

[libvirt] [PATCH RFC v2] qemu: Support numad

by Osier Yang

numad is an user-level daemon that monitors NUMA topology and processes resource consumption to facilitate good NUMA resource alignment of applications/virtual machines to improve performance and minimize cost of remote memory latencies. It provides a pre-placement advisory interface, so significant processes can be pre-bound to nodes with sufficient available resources. More details: http://fedoraproject.org/wiki/Features/numad "numad -w ncpus:memory_amount" is the advisory interface numad provides currently. This patch add the support by introducing a bool XML element: <numatune> <autonuma/> </numatune> If it's specified, the number of vcpus and the current memory amount specified in domain XML will be used for numad command line (numad uses MB for memory amount): numad -w $num_of_vcpus:$current_memory_amount / 1024 The advisory nodeset returned from numad will be used to set domain process CPU affinity then. (e.g. qemuProcessInitCpuAffinity). If the user specifies both CPU affinity policy (e.g. (<vcpu cpuset="1-10,^7,^8">4</vcpu>) and XML indicating to use numad for the advisory nodeset, the specified CPU affinity will be ignored. Only QEMU/KVM and LXC drivers support it now. v1 - v2: * Since Bill Gray says it doesn't matter to use the number of vcpus and current memory amount as numad cmd line argument, though from sementics point of view, what numad expects are physical CPU numbers, let's go this way. v2 dropped XML <cpu required_cpu='4' required_memory='512000'/>, and just a new boolean XML element <autonuma>. Codes are refactored accordingly. * v1 overrides the cpuset specified by <vcpu cpuset='1-10,^7'>2</vcpu>, v2 doesn't do that, but just silently ignored. * xml2xml test is added --- configure.ac | 8 ++ docs/formatdomain.html.in | 11 ++- docs/schemas/domaincommon.rng | 5 + src/conf/domain_conf.c | 124 +++++++++++++------- src/conf/domain_conf.h | 1 + src/lxc/lxc_controller.c | 86 ++++++++++++-- src/qemu/qemu_process.c | 86 ++++++++++++-- .../qemuxml2argv-numatune-autonuma.xml | 32 +++++ tests/qemuxml2xmltest.c | 1 + 9 files changed, 287 insertions(+), 67 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-numatune-autonuma.xml diff --git a/configure.ac b/configure.ac index c9cdd7b..31f0835 100644 --- a/configure.ac +++ b/configure.ac @@ -1445,6 +1445,14 @@ AM_CONDITIONAL([HAVE_NUMACTL], [test "$with_numactl" != "no"]) AC_SUBST([NUMACTL_CFLAGS]) AC_SUBST([NUMACTL_LIBS]) +dnl Do we have numad? +if test "$with_qemu" = "yes"; then + AC_PATH_PROG([NUMAD], [numad], [], [/bin:/usr/bin:/usr/local/bin:$PATH]) + + if test -n "$NUMAD"; then + AC_DEFINE_UNQUOTED([NUMAD],["$NUMAD"], [Location or name of the numad program]) + fi +fi dnl pcap lib LIBPCAP_CONFIG="pcap-config" diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 42f38d3..22872c8 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -505,6 +505,7 @@ ... <numatune> <memory mode="strict" nodeset="1-4,^3"/> + <autonuma/> </numatune> ... </domain> @@ -519,7 +520,7 @@ Since 0.9.3 <dt><code>memory</code></dt> <dd> - The optional <code>memory</code> element specify how to allocate memory + The optional <code>memory</code> element specifies how to allocate memory for the domain process on a NUMA host. It contains two attributes, attribute <code>mode</code> is either 'interleave', 'strict', or 'preferred', @@ -527,6 +528,14 @@ syntax with attribute <code>cpuset</code> of element <code>vcpu</code>. Since 0.9.3 </dd> + <dd> + The optional <code>autonuma</code> element indicates pinning the virtual CPUs + to the advisory nodeset returned by querying "numad" (a system daemon that + monitors NUMA topology and usage). With using this element, the physical CPUs + specified by attribute <code>cpuset</code> (of element <code>vcpu</code>) will + be ignored. + Since 0.9.11 (QEMU/KVM and LXC only) + </dd> </dl> diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index a905457..9066b40 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -549,6 +549,11 @@ </attribute> </element> </optional> + <optional> + <element name="autonuma"> + <empty/> + </element> + </optional> </element> </optional> </interleave> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index b994718..89d00ae 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -7447,7 +7447,6 @@ error: goto cleanup; } - static int virDomainDefMaybeAddController(virDomainDefPtr def, int type, int idx) @@ -7507,6 +7506,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, bool uuid_generated = false; virBitmapPtr bootMap = NULL; unsigned long bootMapSize = 0; + xmlNodePtr cur; if (VIR_ALLOC(def) < 0) { virReportOOMError(); @@ -7776,47 +7776,76 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, VIR_FREE(nodes); /* Extract numatune if exists. */ - if ((n = virXPathNodeSet("./numatune", ctxt, NULL)) < 0) { + if ((n = virXPathNodeSet("./numatune", ctxt, &nodes)) < 0) { virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("cannot extract numatune nodes")); goto error; } + if (n > 1) { + virDomainReportError(VIR_ERR_XML_ERROR, "%s", + _("only one numatune is supported")); + VIR_FREE(nodes); + goto error; + } + if (n) { - tmp = virXPathString("string(./numatune/memory/@nodeset)", ctxt); - if (tmp) { - char *set = tmp; - int nodemasklen = VIR_DOMAIN_CPUMASK_LEN; + cur = nodes[0]->children; + while (cur != NULL) { + if (cur->type == XML_ELEMENT_NODE) { + if ((xmlStrEqual(cur->name, BAD_CAST "memory"))) { + tmp = virXMLPropString(cur, "nodeset"); - if (VIR_ALLOC_N(def->numatune.memory.nodemask, nodemasklen) < 0) { - goto no_memory; - } + if (tmp) { + char *set = tmp; + int nodemasklen = VIR_DOMAIN_CPUMASK_LEN; - /* "nodeset" leads same syntax with "cpuset". */ - if (virDomainCpuSetParse(set, 0, def->numatune.memory.nodemask, - nodemasklen) < 0) - goto error; - VIR_FREE(tmp); - } else { - virDomainReportError(VIR_ERR_INTERNAL_ERROR, - "%s", _("nodeset for NUMA memory tuning must be set")); - goto error; - } + if (VIR_ALLOC_N(def->numatune.memory.nodemask, + nodemasklen) < 0) { + virReportOOMError(); + goto error; + } - tmp = virXPathString("string(./numatune/memory/@mode)", ctxt); - if (tmp) { - if ((def->numatune.memory.mode = - virDomainNumatuneMemModeTypeFromString(tmp)) < 0) { - virDomainReportError(VIR_ERR_INTERNAL_ERROR, - _("Unsupported NUMA memory tuning mode '%s'"), - tmp); - goto error; + /* "nodeset" leads same syntax with "cpuset". */ + if (virDomainCpuSetParse(set, 0, + def->numatune.memory.nodemask, + nodemasklen) < 0) + goto error; + VIR_FREE(tmp); + } else { + virDomainReportError(VIR_ERR_XML_ERROR, "%s", + _("nodeset for NUMA memory " + "tuning must be set")); + goto error; + } + + tmp = virXMLPropString(cur, "mode"); + if (tmp) { + if ((def->numatune.memory.mode = + virDomainNumatuneMemModeTypeFromString(tmp)) < 0) { + virDomainReportError(VIR_ERR_XML_ERROR, + _("Unsupported NUMA memory " + "tuning mode '%s'"), + tmp); + goto error; + } + VIR_FREE(tmp); + } else { + def->numatune.memory.mode = VIR_DOMAIN_NUMATUNE_MEM_STRICT; + } + } else if (xmlStrEqual(cur->name, BAD_CAST "autonuma")) { + def->numatune.autonuma = true; + } else { + virDomainReportError(VIR_ERR_XML_ERROR, + _("unsupported XML element %s"), + (const char *)cur->name); + goto error; + } } - VIR_FREE(tmp); - } else { - def->numatune.memory.mode = VIR_DOMAIN_NUMATUNE_MEM_STRICT; + cur = cur->next; } } + VIR_FREE(nodes); n = virXPathNodeSet("./features/*", ctxt, &nodes); if (n < 0) @@ -12149,23 +12178,30 @@ virDomainDefFormatInternal(virDomainDefPtr def, def->cputune.period || def->cputune.quota) virBufferAddLit(buf, " </cputune>\n"); - if (def->numatune.memory.nodemask) { - const char *mode; - char *nodemask = NULL; - + if (def->numatune.memory.nodemask || + def->numatune.autonuma) { virBufferAddLit(buf, " <numatune>\n"); - nodemask = virDomainCpuSetFormat(def->numatune.memory.nodemask, - VIR_DOMAIN_CPUMASK_LEN); - if (nodemask == NULL) { - virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", - _("failed to format nodeset for NUMA memory tuning")); - goto cleanup; + if (def->numatune.memory.nodemask) { + const char *mode; + char *nodemask = NULL; + + nodemask = virDomainCpuSetFormat(def->numatune.memory.nodemask, + VIR_DOMAIN_CPUMASK_LEN); + if (nodemask == NULL) { + virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to format nodeset for " + "NUMA memory tuning")); + goto cleanup; + } + + mode = virDomainNumatuneMemModeTypeToString(def->numatune.memory.mode); + virBufferAsprintf(buf, " <memory mode='%s' nodeset='%s'/>\n", + mode, nodemask); + VIR_FREE(nodemask); } - mode = virDomainNumatuneMemModeTypeToString(def->numatune.memory.mode); - virBufferAsprintf(buf, " <memory mode='%s' nodeset='%s'/>\n", - mode, nodemask); - VIR_FREE(nodemask); + if (def->numatune.autonuma) + virBufferAddLit(buf, " <autonuma/>\n"); virBufferAddLit(buf, " </numatune>\n"); } diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 87b4103..b15faf0 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -1458,6 +1458,7 @@ struct _virDomainNumatuneDef { int mode; } memory; + bool autonuma; /* Future NUMA tuning related stuff should go here. */ }; diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c index 8f336f5..507b2c6 100644 --- a/src/lxc/lxc_controller.c +++ b/src/lxc/lxc_controller.c @@ -327,6 +327,41 @@ static int lxcSetContainerNUMAPolicy(virDomainDefPtr def) } #endif +#if defined(NUMAD) +static char * +lxcGetNumadAdvice(virDomainDefPtr def) +{ + virCommandPtr cmd = NULL; + char *args = NULL; + char *ret = NULL; + + if (virAsprintf(&args, "%d:%lu", def->vcpus, def->mem.cur_balloon) < 0) { + virReportOOMError(); + goto out; + } + cmd = virCommandNewArgList(NUMAD, "-w", args, NULL); + + virCommandSetOutputBuffer(cmd, &ret); + + if (virCommandRun(cmd, NULL) < 0) { + lxcError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Failed to query numad for the advisory nodeset")); + } + +out: + VIR_FREE(args); + virCommandFree(cmd); + return ret; +} +#else +static char * +lxcGetNumadAdvice(virDomainDefPtr def ATTRIBUTE_UNUSED) +{ + lxcError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("numad is not available on this host")); + return NULL; +} +#endif /* * To be run while still single threaded @@ -355,19 +390,48 @@ static int lxcSetContainerCpuAffinity(virDomainDefPtr def) return -1; } - if (def->cpumask) { - /* XXX why don't we keep 'cpumask' in the libvirt cpumap - * format to start with ?!?! */ - for (i = 0 ; i < maxcpu && i < def->cpumasklen ; i++) - if (def->cpumask[i]) + if (def->numatune.autonuma) { + char *tmp_cpumask = NULL; + char *nodeset = NULL; + + nodeset = lxcGetNumadAdvice(def); + if (!nodeset) + return -1; + + if (VIR_ALLOC_N(tmp_cpumask, VIR_DOMAIN_CPUMASK_LEN) < 0) { + virReportOOMError(); + return -1; + } + + if (virDomainCpuSetParse(nodeset, 0, tmp_cpumask, + VIR_DOMAIN_CPUMASK_LEN) < 0) { + VIR_FREE(tmp_cpumask); + VIR_FREE(nodeset); + return -1; + } + + for (i = 0; i < maxcpu && i < VIR_DOMAIN_CPUMASK_LEN; i++) { + if (tmp_cpumask[i]) VIR_USE_CPU(cpumap, i); + } + + VIR_FREE(tmp_cpumask); + VIR_FREE(nodeset); } else { - /* You may think this is redundant, but we can't assume libvirtd - * itself is running on all pCPUs, so we need to explicitly set - * the spawned LXC instance to all pCPUs if no map is given in - * its config file */ - for (i = 0 ; i < maxcpu ; i++) - VIR_USE_CPU(cpumap, i); + if (def->cpumask) { + /* XXX why don't we keep 'cpumask' in the libvirt cpumap + * format to start with ?!?! */ + for (i = 0 ; i < maxcpu && i < def->cpumasklen ; i++) + if (def->cpumask[i]) + VIR_USE_CPU(cpumap, i); + } else { + /* You may think this is redundant, but we can't assume libvirtd + * itself is running on all pCPUs, so we need to explicitly set + * the spawned LXC instance to all pCPUs if no map is given in + * its config file */ + for (i = 0 ; i < maxcpu ; i++) + VIR_USE_CPU(cpumap, i); + } } /* We are pressuming we are running between fork/exec of LXC diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 7b99814..209403f 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -1633,6 +1633,41 @@ qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm) } #endif +#if defined(NUMAD) +static char * +qemuGetNumadAdvice(virDomainDefPtr def) +{ + virCommandPtr cmd = NULL; + char *args = NULL; + char *output = NULL; + + if (virAsprintf(&args, "%d:%lu", def->vcpus, def->mem.cur_balloon) < 0) { + virReportOOMError(); + goto out; + } + cmd = virCommandNewArgList(NUMAD, "-w", args, NULL); + + virCommandSetOutputBuffer(cmd, &output); + + if (virCommandRun(cmd, NULL) < 0) + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Failed to query numad for the advisory nodeset")); + +out: + VIR_FREE(args); + virCommandFree(cmd); + return output; +} +#else +static char * +qemuGetNumadAdvice(virDomainDefPtr def ATTRIBUTE_UNUSED) +{ + qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("numad is not available on this host")); + return NULL; +} +#endif + /* * To be run between fork/exec of QEMU only */ @@ -1661,19 +1696,48 @@ qemuProcessInitCpuAffinity(virDomainObjPtr vm) return -1; } - if (vm->def->cpumask) { - /* XXX why don't we keep 'cpumask' in the libvirt cpumap - * format to start with ?!?! */ - for (i = 0 ; i < maxcpu && i < vm->def->cpumasklen ; i++) - if (vm->def->cpumask[i]) + if (vm->def->numatune.autonuma) { + char *tmp_cpumask = NULL; + char *nodeset = NULL; + + nodeset = qemuGetNumadAdvice(vm->def); + if (!nodeset) + return -1; + + if (VIR_ALLOC_N(tmp_cpumask, VIR_DOMAIN_CPUMASK_LEN) < 0) { + virReportOOMError(); + return -1; + } + + if (virDomainCpuSetParse(nodeset, 0, tmp_cpumask, + VIR_DOMAIN_CPUMASK_LEN) < 0) { + VIR_FREE(tmp_cpumask); + VIR_FREE(nodeset); + return -1; + } + + for (i = 0; i < maxcpu && i < VIR_DOMAIN_CPUMASK_LEN; i++) { + if (tmp_cpumask[i]) VIR_USE_CPU(cpumap, i); + } + + VIR_FREE(tmp_cpumask); + VIR_FREE(nodeset); } else { - /* You may think this is redundant, but we can't assume libvirtd - * itself is running on all pCPUs, so we need to explicitly set - * the spawned QEMU instance to all pCPUs if no map is given in - * its config file */ - for (i = 0 ; i < maxcpu ; i++) - VIR_USE_CPU(cpumap, i); + if (vm->def->cpumask) { + /* XXX why don't we keep 'cpumask' in the libvirt cpumap + * format to start with ?!?! */ + for (i = 0 ; i < maxcpu && i < vm->def->cpumasklen ; i++) + if (vm->def->cpumask[i]) + VIR_USE_CPU(cpumap, i); + } else { + /* You may think this is redundant, but we can't assume libvirtd + * itself is running on all pCPUs, so we need to explicitly set + * the spawned QEMU instance to all pCPUs if no map is given in + * its config file */ + for (i = 0 ; i < maxcpu ; i++) + VIR_USE_CPU(cpumap, i); + } } /* We are pressuming we are running between fork/exec of QEMU diff --git a/tests/qemuxml2argvdata/qemuxml2argv-numatune-autonuma.xml b/tests/qemuxml2argvdata/qemuxml2argv-numatune-autonuma.xml new file mode 100644 index 0000000..7e50bbd --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-numatune-autonuma.xml @@ -0,0 +1,32 @@ +<domain type='qemu'> + <name>QEMUGuest1</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory>219136</memory> + <currentMemory>219136</currentMemory> + <vcpu current='1'>2</vcpu> + <numatune> + <autonuma/> + </numatune> + <os> + <type arch='i686' machine='pc'>hvm</type> + <boot dev='hd'/> + </os> + <cpu> + <topology sockets='2' cores='1' threads='1'/> + </cpu> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>destroy</on_crash> + <devices> + <emulator>/usr/bin/qemu</emulator> + <disk type='block' device='disk'> + <source dev='/dev/HostVG/QEMUGuest1'/> + <target dev='hda' bus='ide'/> + <address type='drive' controller='0' bus='0' target='0' unit='0'/> + </disk> + <controller type='usb' index='0'/> + <controller type='ide' index='0'/> + <memballoon model='virtio'/> + </devices> +</domain> diff --git a/tests/qemuxml2xmltest.c b/tests/qemuxml2xmltest.c index 03c75f8..3893ece 100644 --- a/tests/qemuxml2xmltest.c +++ b/tests/qemuxml2xmltest.c @@ -199,6 +199,7 @@ mymain(void) DO_TEST("blkiotune"); DO_TEST("blkiotune-device"); DO_TEST("cputune"); + DO_TEST("numatune-autonuma"); DO_TEST("smp"); DO_TEST("lease"); -- 1.7.1

12 years, 9 months

3
4
0 / 0

[libvirt] [PATCH] rpc: allow truncated return for virDomainGetCPUStats

by Eric Blake

The RPC code assumed that the array returned by the driver would be fully populated; that is, ncpus on entry resulted in ncpus * return value on exit. However, while we don't support holes in the middle of ncpus, we do want to permit the case of ncpus on entry being longer than the array returned by the driver (that is, it should be safe for the caller to pass ncpus=128 on entry, and the driver will stop populating the array when it hits max_id). There are now three cases: server 0.9.10 and client 0.9.10 or newer: No impact - there were no hypervisor drivers that supported cpu stats server 0.9.11 or newer and client 0.9.10: if the client calls with ncpus beyond the max, then the rpc call will fail on the client side and disconnect the client, but the server is no worse for the wear server 0.9.11 or newer and client 0.9.11: the server can return a truncated array and the client will do just fine I reproduced the problem by using a host with 2 CPUs, and doing: virsh cpu-stats $dom --start 1 --count 2 * daemon/remote.c (remoteDispatchDomainGetCPUStats): Allow driver to omit tail of array. * src/remote/remote_driver.c (remoteDomainGetCPUStats): Accommodate driver that omits tail of array. --- daemon/remote.c | 10 ++++++++-- src/remote/remote_driver.c | 6 ++++-- 2 files changed, 12 insertions(+), 4 deletions(-) diff --git a/daemon/remote.c b/daemon/remote.c index 74a5f16..39302cc 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -3574,11 +3574,17 @@ remoteDispatchDomainGetCPUStats(virNetServerPtr server ATTRIBUTE_UNUSED, args->flags) < 0) goto cleanup; - percpu_len = ret->params.params_len / args->ncpus; - success: rv = 0; ret->nparams = percpu_len; + if (args->nparams && !(args->flags & VIR_TYPED_PARAM_STRING_OKAY)) { + int i; + + for (i = 0; i < percpu_len; i++) { + if (params[i].type == VIR_TYPED_PARAM_STRING) + ret->nparams--; + } + } cleanup: if (rv < 0) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index 9e74cea..031167a 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -2382,7 +2382,7 @@ static int remoteDomainGetCPUStats(virDomainPtr domain, /* Check the length of the returned list carefully. */ if (ret.params.params_len > nparams * ncpus || (ret.params.params_len && - ret.nparams * ncpus != ret.params.params_len)) { + ((ret.params.params_len % ret.nparams) || ret.nparams > nparams))) { remoteError(VIR_ERR_RPC, "%s", _("remoteDomainGetCPUStats: " "returned number of stats exceeds limit")); @@ -2399,9 +2399,11 @@ static int remoteDomainGetCPUStats(virDomainPtr domain, } /* The remote side did not send back any zero entries, so we have - * to expand things back into a possibly sparse array. + * to expand things back into a possibly sparse array, where the + * tail of the array may be omitted. */ memset(params, 0, sizeof(*params) * nparams * ncpus); + ncpus = ret.params.params_len / ret.nparams; for (cpu = 0; cpu < ncpus; cpu++) { int tmp = nparams; remote_typed_param *stride = &ret.params.params_val[cpu * ret.nparams]; -- 1.7.7.6

12 years, 9 months

2
3
0 / 0

[libvirt] [PATCH] Ensure max_id is initialized in linuxParseCPUmap()

by Daniel P. Berrange

From: "Daniel P. Berrange" <berrange(a)redhat.com> Pushing as a build-breaker fix --- src/nodeinfo.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/nodeinfo.c b/src/nodeinfo.c index 709e94a..61a5925 100644 --- a/src/nodeinfo.c +++ b/src/nodeinfo.c @@ -581,7 +581,7 @@ linuxParseCPUmap(int *max_cpuid, const char *path) { char *map = NULL; char *str = NULL; - int max_id, i; + int max_id = 0, i; if (virFileReadAll(path, 5 * VIR_DOMAIN_CPUMASK_LEN, &str) < 0) { virReportOOMError(); -- 1.7.7.6

12 years, 9 months

1
0
0 / 0

[libvirt] ANNOUNCE: libvirt-glib release 0.0.6

by Daniel P. Berrange

I am pleased to announce that a new release of the libvirt-glib package, version 0.0.6 is now available from ftp://libvirt.org/libvirt/glib/ The packages are GPG signed with Key fingerprint: DAF3 A6FD B26B 6291 2D0E 8E3F BE86 EBB4 1510 4FDF (4096R) New in this release: - Add binding for virDomainBlockResize(): gvir_domain_disk_resize(). - Set correct target node attribute for domain interface. gvir_config_domain_interface_set_ifname() should be setting 'dev' attribute under 'target', not 'device'. - Getter for the associated domain of a domain device. - Getters for GVirConfigDomainInterface attributes. - GVirDomainDevice now has an associated GVirConfigDomainDevice. - Remove now redundant 'path' property from GVirDomainDevice subclasses. - Add gvir_domain_get_devices(). - Empty statistics for user-mode interfaces. One of the limitations of user-mode networking of libvirt is that you can't get statistics for it (not yet, at least). Instead of erroring-out in that case, simply return empty statistics result and spit a debug message. - Fix a GVirStream leak. - Also distribute GNUmakefile, cfg.mk and maint.mk files. libvirt-glib comprises three distinct libraries: - libvirt-glib - Integrate with the GLib event loop and error handling - libvirt-gconfig - Representation of libvirt XML documents as GObjects - libvirt-gobject - Mapping of libvirt APIs into the GObject type system NB: While libvirt aims to be API/ABI stable, for the first few releases, we are *NOT* guaranteeing that libvirt-glib libraries are API/ABI stable. ABI stability will only be guaranteed once the bulk of the APIs have been fleshed out and proved in non-trivial application usage. We anticipate this will be within the next few months in order to line up with Fedora 17. Follow up comments about libvirt-glib should be directed to the regular libvir-list redhat com development list. Thanks to all the people involved in contributing to this release. Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|

12 years, 9 months

1
0
0 / 0

[libvirt] [PATCH] docs: Fix typo

by Osier Yang

It used "&lt" for ">", reported by Kyla Zhang <weizhan(a)redhat.com> -- Pushed under trivial rule. --- docs/formatdomain.html.in | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 6434ae5..42f38d3 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1096,9 +1096,9 @@ </disk> <disk type='block' device='lun'> <driver name='qemu' type='raw'/> - <source dev='/dev/sda'/< - <target dev='sda' bus='scsi'/< - <address type='drive' controller='0' bus='0' target='3' unit='0'/< + <source dev='/dev/sda'/> + <target dev='sda' bus='scsi'/> + <address type='drive' controller='0' bus='0' target='3' unit='0'/> </disk> </devices> ...</pre> -- 1.7.1

12 years, 9 months

1
0
0 / 0

[libvirt] [PATCH RFC]: Support numad

by Osier Yang

numad is an user-level daemon that monitors NUMA topology and processes resource consumption to facilitate good NUMA resource alignment of applications/virtual machines to improve performance and minimize cost of remote memory latencies. It provides a pre-placement advisory interface, so significant processes can be pre-bound to nodes with sufficient available resources. More details: http://fedoraproject.org/wiki/Features/numad "numad -w ncpus:memory_amount" is the advisory interface numad provides currently. This patch add the support by introducing new XML like: <numatune> <cpu required_cpus="4" required_memory="524288"/> </numatune> And the corresponding numad command line will be: numad -w 4:500 The advisory nodeset returned from numad will be used to set domain process CPU affinity then. (e.g. qemuProcessInitCpuAffinity). If the user specifies both CPU affinity policy (e.g. (<vcpu cpuset="1-10,^7,^8">4</vcpu>) and XML indicating to use numad for the advisory nodeset, the specified CPU affinity will be overridden by the nodeset returned from numad. If no XML to specify the CPU affinity policy, and XML indicating to use numad is specified, the returned nodeset will be printed in <cpu cpuset="$nodeset_from_numad"/>4</vcpu>. Only QEMU/KVM and LXC drivers support it now. --- configure.ac | 8 +++ docs/formatdomain.html.in | 18 ++++++- docs/schemas/domaincommon.rng | 12 ++++ src/conf/domain_conf.c | 125 +++++++++++++++++++++++++++++++---------- src/conf/domain_conf.h | 5 ++ src/lxc/lxc_controller.c | 98 ++++++++++++++++++++++++++++---- src/qemu/qemu_process.c | 99 +++++++++++++++++++++++++++++---- 7 files changed, 311 insertions(+), 54 deletions(-) diff --git a/configure.ac b/configure.ac index c9cdd7b..31f0835 100644 --- a/configure.ac +++ b/configure.ac @@ -1445,6 +1445,14 @@ AM_CONDITIONAL([HAVE_NUMACTL], [test "$with_numactl" != "no"]) AC_SUBST([NUMACTL_CFLAGS]) AC_SUBST([NUMACTL_LIBS]) +dnl Do we have numad? +if test "$with_qemu" = "yes"; then + AC_PATH_PROG([NUMAD], [numad], [], [/bin:/usr/bin:/usr/local/bin:$PATH]) + + if test -n "$NUMAD"; then + AC_DEFINE_UNQUOTED([NUMAD],["$NUMAD"], [Location or name of the numad program]) + fi +fi dnl pcap lib LIBPCAP_CONFIG="pcap-config" diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 6fcca94..d8e70a6 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -505,6 +505,7 @@ ... <numatune> <memory mode="strict" nodeset="1-4,^3"/> + <cpu required_cpus="3" required_memory="524288"/> </numatune> ... </domain> @@ -519,7 +520,7 @@ Since 0.9.3 <dt><code>memory</code></dt> <dd> - The optional <code>memory</code> element specify how to allocate memory + The optional <code>memory</code> element specifies how to allocate memory for the domain process on a NUMA host. It contains two attributes, attribute <code>mode</code> is either 'interleave', 'strict', or 'preferred', @@ -527,6 +528,21 @@ syntax with attribute <code>cpuset</code> of element <code>vcpu</code>. Since 0.9.3 </dd> + <dd> + The optional <code>cpu</code> element indicates pinning the virtual CPUs + to the nodeset returned by querying "numad" (a system daemon that monitors + NUMA topology and usage). It has two attributes, attribute + <code>required_cpus</code> specifies the number of physical CPUs the guest + process want to use. And the optional attribute <code>required_memory</code> + specifies the amount of free memory the guest process want to see on a node, + "numad" will pick the physical CPUs on the node which has enough free + memory of amount specified by <code>required_memory</code>. + + NB, with using this element, the physical CPUs specified by attribute + <code>cpuset</code> (of element <code>vcpu</code>) will be overridden by the + nodeset returned from "numad". + Since 0.9.11 (QEMU/KVM and LXC only) + </dd> </dl> diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index 3908733..d0f443d 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -549,6 +549,18 @@ </attribute> </element> </optional> + <optional> + <element name="cpu"> + <attribute name="required_cpu"> + <ref name="countCPU"/> + </attribute> + <optional> + <attribute name="required_memory"> + <ref name="memoryKB"/> + </attribute> + </optional> + </element> + </optional> </element> </optional> </interleave> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index f9654f1..aa03c05 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -7125,7 +7125,6 @@ error: goto cleanup; } - static int virDomainDefMaybeAddController(virDomainDefPtr def, int type, int idx) @@ -7185,6 +7184,7 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, bool uuid_generated = false; virBitmapPtr bootMap = NULL; unsigned long bootMapSize = 0; + xmlNodePtr cur; if (VIR_ALLOC(def) < 0) { virReportOOMError(); @@ -7454,47 +7454,100 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, VIR_FREE(nodes); /* Extract numatune if exists. */ - if ((n = virXPathNodeSet("./numatune", ctxt, NULL)) < 0) { + if ((n = virXPathNodeSet("./numatune", ctxt, &nodes)) < 0) { virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("cannot extract numatune nodes")); goto error; } + if (n > 1) { + virDomainReportError(VIR_ERR_XML_ERROR, "%s", + _("only one numatune is supported")); + VIR_FREE(nodes); + goto error; + } + if (n) { - tmp = virXPathString("string(./numatune/memory/@nodeset)", ctxt); - if (tmp) { - char *set = tmp; - int nodemasklen = VIR_DOMAIN_CPUMASK_LEN; + cur = nodes[0]->children; + while (cur != NULL) { + if (cur->type == XML_ELEMENT_NODE) { + if ((xmlStrEqual(cur->name, BAD_CAST "memory"))) { + tmp = virXMLPropString(cur, "nodeset"); - if (VIR_ALLOC_N(def->numatune.memory.nodemask, nodemasklen) < 0) { - goto no_memory; - } + if (tmp) { + char *set = tmp; + int nodemasklen = VIR_DOMAIN_CPUMASK_LEN; - /* "nodeset" leads same syntax with "cpuset". */ - if (virDomainCpuSetParse(set, 0, def->numatune.memory.nodemask, - nodemasklen) < 0) - goto error; - VIR_FREE(tmp); - } else { - virDomainReportError(VIR_ERR_INTERNAL_ERROR, - "%s", _("nodeset for NUMA memory tuning must be set")); - goto error; - } + if (VIR_ALLOC_N(def->numatune.memory.nodemask, + nodemasklen) < 0) { + virReportOOMError(); + goto error; + } - tmp = virXPathString("string(./numatune/memory/@mode)", ctxt); - if (tmp) { - if ((def->numatune.memory.mode = - virDomainNumatuneMemModeTypeFromString(tmp)) < 0) { - virDomainReportError(VIR_ERR_INTERNAL_ERROR, - _("Unsupported NUMA memory tuning mode '%s'"), - tmp); - goto error; + /* "nodeset" leads same syntax with "cpuset". */ + if (virDomainCpuSetParse(set, 0, + def->numatune.memory.nodemask, + nodemasklen) < 0) + goto error; + VIR_FREE(tmp); + } else { + virDomainReportError(VIR_ERR_XML_ERROR, "%s", + _("nodeset for NUMA memory " + "tuning must be set")); + goto error; + } + + tmp = virXMLPropString(cur, "mode"); + if (tmp) { + if ((def->numatune.memory.mode = + virDomainNumatuneMemModeTypeFromString(tmp)) < 0) { + virDomainReportError(VIR_ERR_XML_ERROR, + _("Unsupported NUMA memory " + "tuning mode '%s'"), + tmp); + goto error; + } + VIR_FREE(tmp); + } else { + def->numatune.memory.mode = VIR_DOMAIN_NUMATUNE_MEM_STRICT; + } + } else if (xmlStrEqual(cur->name, BAD_CAST "cpu")) { + char *req_cpus = NULL; + char *req_memory = NULL; + req_cpus = virXMLPropString(cur, "required_cpus"); + req_memory = virXMLPropString(cur, "required_memory"); + + if (req_cpus && + virStrToLong_ui(req_cpus, NULL, 10, + &def->numatune.cpu.required_cpus) < 0) { + virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Cannot parse <cpu> 'required_cpus'" + " attribute")); + goto error; + } + + if (req_memory && + virStrToLong_ul(req_memory, NULL, 10, + &def->numatune.cpu.required_memory) < 0) { + virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Cannot parse <cpu> 'required_memory'" + " attribute")); + goto error; + } + + VIR_FREE(req_cpus); + VIR_FREE(req_memory); + } else { + virDomainReportError(VIR_ERR_XML_ERROR, + _("unsupported XML element %s"), + (const char *)cur->name); + goto error; + } } - VIR_FREE(tmp); - } else { - def->numatune.memory.mode = VIR_DOMAIN_NUMATUNE_MEM_STRICT; + cur = cur->next; } } + VIR_FREE(nodes); n = virXPathNodeSet("./features/*", ctxt, &nodes); if (n < 0) @@ -11761,7 +11814,8 @@ virDomainDefFormatInternal(virDomainDefPtr def, def->cputune.period || def->cputune.quota) virBufferAddLit(buf, " </cputune>\n"); - if (def->numatune.memory.nodemask) { + if (def->numatune.memory.nodemask || + def->numatune.cpu.required_cpus) { const char *mode; char *nodemask = NULL; @@ -11778,6 +11832,15 @@ virDomainDefFormatInternal(virDomainDefPtr def, virBufferAsprintf(buf, " <memory mode='%s' nodeset='%s'/>\n", mode, nodemask); VIR_FREE(nodemask); + + if (def->numatune.cpu.required_cpus) + virBufferAsprintf(buf, " <cpu required_cpus='%d' ", + def->numatune.cpu.required_cpus); + + if (def->numatune.cpu.required_memory) + virBufferAsprintf(buf, "required_memory='%lu'/>\n", + def->numatune.cpu.required_memory); + virBufferAddLit(buf, " </numatune>\n"); } diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 596be4d..1284599 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -1416,6 +1416,11 @@ struct _virDomainNumatuneDef { int mode; } memory; + struct { + unsigned int required_cpus; + unsigned long required_memory; + } cpu; + /* Future NUMA tuning related stuff should go here. */ }; diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c index 8f336f5..ec6434d 100644 --- a/src/lxc/lxc_controller.c +++ b/src/lxc/lxc_controller.c @@ -327,6 +327,47 @@ static int lxcSetContainerNUMAPolicy(virDomainDefPtr def) } #endif +#if defined(NUMAD) +static char * +lxcGetNumadAdvice(unsigned int req_cpus, + unsigned long req_memory) { + virCommandPtr cmd = NULL; + char *reqs = NULL; + char *ret = NULL; + + /* numad uses "MB" for memory. */ + if (req_memory) { + req_memory = req_memory / 1024; + if (virAsprintf(&reqs, "%d:%lu", req_cpus, req_memory) < 0) { + virReportOOMError(); + goto out; + } + cmd = virCommandNewArgList(NUMAD, "-w", reqs, NULL); + } else { + cmd = virCommandNewArgList(NUMAD, "-w", "%d", req_cpus, NULL); + } + + virCommandSetOutputBuffer(cmd, &ret); + + if (virCommandRun(cmd, NULL) < 0) { + lxcError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Failed to query numad for the advisory nodeset")); + } + +out: + VIR_FREE(reqs); + virCommandFree(cmd); + return ret; +} +#else +static char * +lxcGetNumadAdvice(unsigned int req_cpus ATTRIBUTE_UNUSED, + unsigned long req_memory ATTRIBUTE_UNUSED) { + lxcError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("numad is not available on this host")); + return NULL; +} +#endif /* * To be run while still single threaded @@ -355,19 +396,54 @@ static int lxcSetContainerCpuAffinity(virDomainDefPtr def) return -1; } - if (def->cpumask) { - /* XXX why don't we keep 'cpumask' in the libvirt cpumap - * format to start with ?!?! */ - for (i = 0 ; i < maxcpu && i < def->cpumasklen ; i++) - if (def->cpumask[i]) + /* def->cpumask will be overridden by the nodeset + * suggested by numad if it's specified. + */ + if (def->numatune.cpu.required_cpus) { + char *tmp_cpumask = NULL; + char *nodeset = NULL; + + nodeset = lxcGetNumadAdvice(def->numatune.cpu.required_cpus, + def->numatune.cpu.required_memory); + if (!nodeset) + return -1; + + if (VIR_ALLOC_N(tmp_cpumask, VIR_DOMAIN_CPUMASK_LEN) < 0) { + virReportOOMError(); + return -1; + } + + if (virDomainCpuSetParse(nodeset, 0, tmp_cpumask, + VIR_DOMAIN_CPUMASK_LEN) < 0) { + VIR_FREE(tmp_cpumask); + VIR_FREE(nodeset); + return -1; + } + + for (i = 0; i < maxcpu && i < VIR_DOMAIN_CPUMASK_LEN; i++) { + if (tmp_cpumask[i]) VIR_USE_CPU(cpumap, i); + } + + /* Update def->cpumask */ + VIR_FREE(def->cpumask); + def->cpumask = tmp_cpumask; + VIR_FREE(nodeset); } else { - /* You may think this is redundant, but we can't assume libvirtd - * itself is running on all pCPUs, so we need to explicitly set - * the spawned LXC instance to all pCPUs if no map is given in - * its config file */ - for (i = 0 ; i < maxcpu ; i++) - VIR_USE_CPU(cpumap, i); + if (def->cpumask) { + /* XXX why don't we keep 'cpumask' in the libvirt cpumap + * format to start with ?!?! */ + for (i = 0 ; i < maxcpu && i < def->cpumasklen ; i++) + if (def->cpumask[i]) + VIR_USE_CPU(cpumap, i); + } else { + /* You may think this is redundant, but we can't assume libvirtd + * itself is running on all pCPUs, so we need to explicitly set + * the spawned LXC instance to all pCPUs if no map is given in + * its config file */ + for (i = 0 ; i < maxcpu ; i++) + VIR_USE_CPU(cpumap, i); + } } /* We are pressuming we are running between fork/exec of LXC diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 41218de..eb9f8f1 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -1633,6 +1633,48 @@ qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm) } #endif +#if defined(NUMAD) +static char * +qemuGetNumadAdvice(unsigned int req_cpus, + unsigned long req_memory) { + virCommandPtr cmd = NULL; + char *reqs = NULL; + char *output = NULL; + + /* numad uses "MB" for memory. */ + if (req_memory) { + req_memory = req_memory / 1024; + if (virAsprintf(&reqs, "%d:%lu", req_cpus, req_memory) < 0) { + virReportOOMError(); + goto out; + } + + cmd = virCommandNewArgList(NUMAD, "-w", reqs, NULL); + } else { + cmd = virCommandNewArgList(NUMAD, "-w", "%u", req_cpus, NULL); + } + + virCommandSetOutputBuffer(cmd, &output); + + if (virCommandRun(cmd, NULL) < 0) + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Failed to query numad for the advisory nodeset")); + +out: + VIR_FREE(reqs); + virCommandFree(cmd); + return output; +} +#else +static char * +qemuGetNumadAdvice(unsigned int req_cpus ATTRIBUTE_UNUSED, + unsigned long req_memory ATTRIBUTE_UNUSED) { + qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("numad is not available on this host")); + return NULL; +} +#endif + /* * To be run between fork/exec of QEMU only */ @@ -1661,19 +1703,54 @@ qemuProcessInitCpuAffinity(virDomainObjPtr vm) return -1; } - if (vm->def->cpumask) { - /* XXX why don't we keep 'cpumask' in the libvirt cpumap - * format to start with ?!?! */ - for (i = 0 ; i < maxcpu && i < vm->def->cpumasklen ; i++) - if (vm->def->cpumask[i]) + /* vm->def->cpumask will be overridden by the nodeset + * suggested by numad if it's specified. + */ + if (vm->def->numatune.cpu.required_cpus) { + char *tmp_cpumask = NULL; + char *nodeset = NULL; + + nodeset = qemuGetNumadAdvice(vm->def->numatune.cpu.required_cpus, + vm->def->numatune.cpu.required_memory); + if (!nodeset) + return -1; + + if (VIR_ALLOC_N(tmp_cpumask, VIR_DOMAIN_CPUMASK_LEN) < 0) { + virReportOOMError(); + return -1; + } + + if (virDomainCpuSetParse(nodeset, 0, tmp_cpumask, + VIR_DOMAIN_CPUMASK_LEN) < 0) { + VIR_FREE(tmp_cpumask); + VIR_FREE(nodeset); + return -1; + } + + for (i = 0; i < maxcpu && i < VIR_DOMAIN_CPUMASK_LEN; i++) { + if (tmp_cpumask[i]) VIR_USE_CPU(cpumap, i); + } + + /* Update vm->def->cpumask */ + VIR_FREE(vm->def->cpumask); + vm->def->cpumask = tmp_cpumask; + VIR_FREE(nodeset); } else { - /* You may think this is redundant, but we can't assume libvirtd - * itself is running on all pCPUs, so we need to explicitly set - * the spawned QEMU instance to all pCPUs if no map is given in - * its config file */ - for (i = 0 ; i < maxcpu ; i++) - VIR_USE_CPU(cpumap, i); + if (vm->def->cpumask) { + /* XXX why don't we keep 'cpumask' in the libvirt cpumap + * format to start with ?!?! */ + for (i = 0 ; i < maxcpu && i < vm->def->cpumasklen ; i++) + if (vm->def->cpumask[i]) + VIR_USE_CPU(cpumap, i); + } else { + /* You may think this is redundant, but we can't assume libvirtd + * itself is running on all pCPUs, so we need to explicitly set + * the spawned QEMU instance to all pCPUs if no map is given in + * its config file */ + for (i = 0 ; i < maxcpu ; i++) + VIR_USE_CPU(cpumap, i); + } } /* We are pressuming we are running between fork/exec of QEMU -- 1.7.7.3

12 years, 9 months

5
8
0 / 0

[libvirt] [PATCH] add screendump async to qemu

by Alon Levy

RHBZ: 800338 Adds a new capability to qemu, QEMU_CAPS_SCREENDUMP_ASYNC, available if the qmp command "screendump-async" exists. If that cap exists qemuDomainScreenshot uses it. The implementation consists of a hash from filename to struct holding the stream and temporary fd. The fd is closed and the stream is written to (in reverse order) by the completion callback, qemuProcessScreenshotComplete. Note: in qemuDomainScreenshot I don't check for an existing entry in the screenshots hash table because we the key is a temporary filename, produced by mkstemp, and it's only unlinked at qemuProcessScreenshotComplete. For testing you need to apply the following patches (they are still pending review on qemu-devel): http://patchwork.ozlabs.org/patch/144706/ http://patchwork.ozlabs.org/patch/144705/ http://patchwork.ozlabs.org/patch/144704/ Signed-off-by: Alon Levy <alevy(a)redhat.com> --- src/qemu/qemu_capabilities.c | 1 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_domain.c | 6 ++++++ src/qemu/qemu_domain.h | 12 ++++++++++++ src/qemu/qemu_driver.c | 42 +++++++++++++++++++++++++++++++++++------- src/qemu/qemu_monitor.c | 26 ++++++++++++++++++++++++++ src/qemu/qemu_monitor.h | 8 ++++++++ src/qemu/qemu_monitor_json.c | 39 +++++++++++++++++++++++++++++++++++++++ src/qemu/qemu_monitor_json.h | 3 +++ src/qemu/qemu_process.c | 29 +++++++++++++++++++++++++++++ 10 files changed, 160 insertions(+), 7 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 64a4546..57771ff 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -154,6 +154,7 @@ VIR_ENUM_IMPL(qemuCaps, QEMU_CAPS_LAST, "drive-iotune", /* 85 */ "system_wakeup", "scsi-disk.channel", + "screendump-async", ); struct qemu_feature_flags { diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index db584ce..24d620d 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -122,6 +122,7 @@ enum qemuCapsFlags { QEMU_CAPS_DRIVE_IOTUNE = 85, /* -drive bps= and friends */ QEMU_CAPS_WAKEUP = 86, /* system_wakeup monitor command */ QEMU_CAPS_SCSI_DISK_CHANNEL = 87, /* Is scsi-disk.channel available? */ + QEMU_CAPS_SCREENDUMP_ASYNC = 88, /* screendump-async qmp command */ QEMU_CAPS_LAST, /* this must always be the last item */ }; diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 2fed91e..acf56c4 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -181,6 +181,10 @@ qemuDomainObjFreeJob(qemuDomainObjPrivatePtr priv) ignore_value(virCondDestroy(&priv->job.asyncCond)); } +static void +freeScreenshot(void *payload, const void *name ATTRIBUTE_UNUSED) { + VIR_FREE(payload); +} static void *qemuDomainObjPrivateAlloc(void) { @@ -196,6 +200,8 @@ static void *qemuDomainObjPrivateAlloc(void) goto error; priv->migMaxBandwidth = QEMU_DOMAIN_DEFAULT_MIG_BANDWIDTH_MAX; + priv->screenshots = virHashCreate(QEMU_DOMAIN_SCREENSHOTS_CONCURRENT_MAX, + freeScreenshot); return priv; diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h index 1333d8c..15721ec 100644 --- a/src/qemu/qemu_domain.h +++ b/src/qemu/qemu_domain.h @@ -40,6 +40,8 @@ # define QEMU_DOMAIN_DEFAULT_MIG_BANDWIDTH_MAX 32 +# define QEMU_DOMAIN_SCREENSHOTS_CONCURRENT_MAX 16 + # define JOB_MASK(job) (1 << (job - 1)) # define DEFAULT_JOB_MASK \ (JOB_MASK(QEMU_JOB_QUERY) | \ @@ -91,6 +93,14 @@ struct qemuDomainJobObj { virDomainJobInfo info; /* Async job progress data */ }; +struct _qemuScreenshotAsync { + virStreamPtr stream; /* stream to write results to */ + const char *filename; /* temporary file to read results from */ + int fd; /* handle to open temporary file */ +}; +typedef struct _qemuScreenshotAsync qemuScreenshotAsync; +typedef qemuScreenshotAsync *qemuScreenshotAsyncPtr; + typedef struct _qemuDomainPCIAddressSet qemuDomainPCIAddressSet; typedef qemuDomainPCIAddressSet *qemuDomainPCIAddressSetPtr; @@ -130,6 +140,8 @@ struct _qemuDomainObjPrivate { char *origname; virConsolesPtr cons; + + virHashTablePtr screenshots; }; struct qemuDomainWatchdogEvent diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 733df0a..e397bc3 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -3135,6 +3135,7 @@ qemuDomainScreenshot(virDomainPtr dom, int tmp_fd = -1; char *ret = NULL; bool unlink_tmp = false; + qemuScreenshotAsync *screenshot; virCheckFlags(0, NULL); @@ -3184,9 +3185,34 @@ qemuDomainScreenshot(virDomainPtr dom, virSecurityManagerSetSavedStateLabel(qemu_driver->securityManager, vm->def, tmp); qemuDomainObjEnterMonitor(driver, vm); - if (qemuMonitorScreendump(priv->mon, tmp) < 0) { - qemuDomainObjExitMonitor(driver, vm); - goto endjob; + if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_SCREENDUMP_ASYNC)) { + if (virHashSize(priv->screenshots) >= + QEMU_DOMAIN_SCREENSHOTS_CONCURRENT_MAX) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("too many ongoing screenshots")); + goto endjob; + } + if (VIR_ALLOC(screenshot) < 0) { + qemuReportError(VIR_ERR_NO_MEMORY, "%s", _("out of memory")); + goto endjob; + } + screenshot->fd = tmp_fd; + screenshot->filename = tmp; + screenshot->stream = st; + virHashAddEntry(priv->screenshots, tmp, screenshot); + if (qemuMonitorScreendumpAsync(priv->mon, tmp) < 0) { + qemuDomainObjExitMonitor(driver, vm); + goto endjob; + } + /* string and fd are freed by qmp event callback */ + tmp = NULL; + tmp_fd = -1; + unlink_tmp = false; + } else { + if (qemuMonitorScreendump(priv->mon, tmp) < 0) { + qemuDomainObjExitMonitor(driver, vm); + goto endjob; + } } qemuDomainObjExitMonitor(driver, vm); @@ -3195,10 +3221,12 @@ qemuDomainScreenshot(virDomainPtr dom, goto endjob; } - if (virFDStreamOpenFile(st, tmp, 0, 0, O_RDONLY) < 0) { - qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", - _("unable to open stream")); - goto endjob; + if (!qemuCapsGet(priv->qemuCaps, QEMU_CAPS_SCREENDUMP_ASYNC)) { + if (virFDStreamOpenFile(st, tmp, 0, 0, O_RDONLY) < 0) { + qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", + _("unable to open stream")); + goto endjob; + } } ret = strdup("image/x-portable-pixmap"); diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c index 1da73f6..0df63e7 100644 --- a/src/qemu/qemu_monitor.c +++ b/src/qemu/qemu_monitor.c @@ -1054,6 +1054,18 @@ int qemuMonitorEmitBlockJob(qemuMonitorPtr mon, } +int qemuMonitorEmitScreenDumpComplete(qemuMonitorPtr mon, + const char *filename) +{ + int ret = -1; + VIR_DEBUG("mon=%p", mon); + + QEMU_MONITOR_CALLBACK(mon, ret, domainScreenshotComplete, mon->vm, + filename); + return ret; +} + + int qemuMonitorSetCapabilities(qemuMonitorPtr mon, virBitmapPtr qemuCaps) @@ -2710,6 +2722,20 @@ int qemuMonitorScreendump(qemuMonitorPtr mon, return ret; } +int qemuMonitorScreendumpAsync(qemuMonitorPtr mon, + const char *file) +{ + VIR_DEBUG("mon=%p, file=%s", mon, file); + + if (!mon) { + qemuReportError(VIR_ERR_INVALID_ARG,"%s", + _("monitor must not be NULL")); + return -1; + } + + return qemuMonitorJSONScreendumpAsync(mon, file); +} + int qemuMonitorBlockJob(qemuMonitorPtr mon, const char *device, const char *base, diff --git a/src/qemu/qemu_monitor.h b/src/qemu/qemu_monitor.h index b1c956c..abe977c 100644 --- a/src/qemu/qemu_monitor.h +++ b/src/qemu/qemu_monitor.h @@ -124,6 +124,9 @@ struct _qemuMonitorCallbacks { const char *diskAlias, int type, int status); + int (*domainScreenshotComplete)(qemuMonitorPtr mon, + virDomainObjPtr vm, + const char *filename); }; @@ -195,6 +198,8 @@ int qemuMonitorEmitBlockJob(qemuMonitorPtr mon, const char *diskAlias, int type, int status); +int qemuMonitorEmitScreenDumpComplete(qemuMonitorPtr mon, + const char *filename); @@ -505,6 +510,9 @@ int qemuMonitorInjectNMI(qemuMonitorPtr mon); int qemuMonitorScreendump(qemuMonitorPtr mon, const char *file); +int qemuMonitorScreendumpAsync(qemuMonitorPtr mon, + const char *file); + int qemuMonitorSendKey(qemuMonitorPtr mon, unsigned int holdtime, unsigned int *keycodes, diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index dc67b4b..79ec6ba 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -59,6 +59,7 @@ static void qemuMonitorJSONHandleVNCConnect(qemuMonitorPtr mon, virJSONValuePtr static void qemuMonitorJSONHandleVNCInitialize(qemuMonitorPtr mon, virJSONValuePtr data); static void qemuMonitorJSONHandleVNCDisconnect(qemuMonitorPtr mon, virJSONValuePtr data); static void qemuMonitorJSONHandleBlockJob(qemuMonitorPtr mon, virJSONValuePtr data); +static void qemuMonitorJSONHandleScreenDumpComplete(qemuMonitorPtr mon, virJSONValuePtr data); static struct { const char *type; @@ -75,6 +76,7 @@ static struct { { "VNC_INITIALIZED", qemuMonitorJSONHandleVNCInitialize, }, { "VNC_DISCONNECTED", qemuMonitorJSONHandleVNCDisconnect, }, { "BLOCK_JOB_COMPLETED", qemuMonitorJSONHandleBlockJob, }, + { "SCREEN_DUMP_COMPLETE", qemuMonitorJSONHandleScreenDumpComplete, }, }; @@ -725,6 +727,16 @@ out: qemuMonitorEmitBlockJob(mon, device, type, status); } +static void qemuMonitorJSONHandleScreenDumpComplete(qemuMonitorPtr mon, + virJSONValuePtr data) +{ + const char *filename; + + if ((filename = virJSONValueObjectGetString(data, "filename")) == NULL) { + VIR_WARN("missing filename in screen dump complete event"); + } + qemuMonitorEmitScreenDumpComplete(mon, filename); +} int qemuMonitorJSONHumanCommandWithFd(qemuMonitorPtr mon, @@ -836,6 +848,9 @@ qemuMonitorJSONCheckCommands(qemuMonitorPtr mon, if (STREQ(name, "system_wakeup")) qemuCapsSet(qemuCaps, QEMU_CAPS_WAKEUP); + + if (STREQ(name, "screendump-async")) + qemuCapsSet(qemuCaps, QEMU_CAPS_SCREENDUMP_ASYNC); } ret = 0; @@ -3135,6 +3150,30 @@ int qemuMonitorJSONScreendump(qemuMonitorPtr mon, return ret; } +int qemuMonitorJSONScreendumpAsync(qemuMonitorPtr mon, + const char *file) +{ + int ret; + virJSONValuePtr cmd, reply = NULL; + + cmd = qemuMonitorJSONMakeCommand("screendump-async", + "s:filename", file, + NULL); + + if (!cmd) + return -1; + + ret = qemuMonitorJSONCommand(mon, cmd, &reply); + + if (ret == 0) + ret = qemuMonitorJSONCheckError(cmd, reply); + + virJSONValueFree(cmd); + virJSONValueFree(reply); + return ret; +} + + static int qemuMonitorJSONGetBlockJobInfoOne(virJSONValuePtr entry, const char *device, virDomainBlockJobInfoPtr info) diff --git a/src/qemu/qemu_monitor_json.h b/src/qemu/qemu_monitor_json.h index 0932a2c..74acce1 100644 --- a/src/qemu/qemu_monitor_json.h +++ b/src/qemu/qemu_monitor_json.h @@ -244,6 +244,9 @@ int qemuMonitorJSONSendKey(qemuMonitorPtr mon, int qemuMonitorJSONScreendump(qemuMonitorPtr mon, const char *file); +int qemuMonitorJSONScreendumpAsync(qemuMonitorPtr mon, + const char *file); + int qemuMonitorJSONBlockJob(qemuMonitorPtr mon, const char *device, const char *base, diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 7b99814..7985a37 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -47,6 +47,7 @@ #include "datatypes.h" #include "logging.h" +#include "fdstream.h" #include "virterror_internal.h" #include "memory.h" #include "hooks.h" @@ -918,6 +919,33 @@ qemuProcessHandleBlockJob(qemuMonitorPtr mon ATTRIBUTE_UNUSED, } static int +qemuProcessScreenshotComplete(qemuMonitorPtr mon ATTRIBUTE_UNUSED, + virDomainObjPtr vm, + const char *filename) +{ + qemuDomainObjPrivatePtr priv = vm->privateData; + qemuScreenshotAsyncPtr screenshot; + int ret = 0; + + if ((screenshot = virHashLookup(priv->screenshots, filename)) == NULL) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("got screendump completion event for wrong filename")); + ret = -1; + goto end; + } + if (virFDStreamOpenFile(screenshot->stream, filename, 0, 0, O_RDONLY) < 0) { + qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", + _("unable to open stream")); + ret = -1; + } +end: + VIR_FORCE_CLOSE(screenshot->fd); + VIR_FREE(screenshot->filename); + virHashRemoveEntry(priv->screenshots, filename); + return ret; +} + +static int qemuProcessHandleGraphics(qemuMonitorPtr mon ATTRIBUTE_UNUSED, virDomainObjPtr vm, int phase, @@ -1034,6 +1062,7 @@ static qemuMonitorCallbacks monitorCallbacks = { .domainIOError = qemuProcessHandleIOError, .domainGraphics = qemuProcessHandleGraphics, .domainBlockJob = qemuProcessHandleBlockJob, + .domainScreenshotComplete = qemuProcessScreenshotComplete, }; static int -- 1.7.9.1

12 years, 9 months

2
2
0 / 0

[libvirt] [PATCH 0/4 v3] Support mac and port profile for <interface type='hostdev'>

by Roopa Prabhu

v3: Changes include: - Review comments from Laine - rebased with latest upstream v2: changes include: - feedback from stefan for 802.1Qbg. Code now prints an error if virtualport is specified for 802.1Qbg on an interface of type hostdev - feedback from laine for non-sriov devices. Interface type hostdev for non-sriov devices is not supported. v1: https://www.redhat.com/archives/libvir-list/2012-March/msg00015.html This patch series is based on laines patches to support <interface type='hostdev'>. https://www.redhat.com/archives/libvir-list/2012-February/msg01126.html It support to set mac and port profile on an interface of type hostdev. * If virtualport is specified, the existing virtual port functions are called to set mac, vlan and port profile. * If virtualport is not specified and device is a sriov virtual function, - mac is set using IFLA_VF_MAC * If virtualport is not specified and device is a non-sriov virtual function, - mac is set using existing SIOCGIFHWADDR (This requires that the netdev be present on the host before starting the VM) This series implements the below : 01/4 pci: Add two new pci util pciDeviceGetVirtualFunctionInfo and pciConfigAddressToSysfsFile 02/4 virtnetdev: Add support functions for mac and portprofile associations on a hostdev 03/4 virnetdevvportprofile: Changes to support portprofiles for hostdevs 04/4 qemu_hostdev: Add support to install port profile and mac address on hostdev Stefan Berger is CC'ed for 802.1Qbg changes in patch 03/4. Current code for 802.1Qbg uses macvtap ifname. And for network interfaces with type=hostdev a macvtap ifname does not exist. This patch just adds a null check for ifname in 802.1Qbg port profile handling code.

12 years, 9 months

2
11
0 / 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Devel March 2012