- Devel - libvirt List Archives

[libvirt] [PATCH] cpustat: fix regression when cpus are offline
by Eric Blake 26 Oct '12

26 Oct '12

It turns out that the cpuacct results properly account for offline cpus, and always returns results for every possible cpu, not just the online ones. So there is no need to check the map of online cpus in the first place, merely only a need to know the maximum possible cpu. Meanwhile, virNodeGetCPUBitmap had a subtle change from returning the maximum id to instead returning the width of the bitmap (one larger than the maximum id), which made this code encounter some off-by-one logic leading to bad error messages when a cpu was offline: $ virsh cpu-stats dom error: Failed to virDomainGetCPUStats() error: An error occurred, but the cause is unknown * src/qemu/qemu_driver.c (qemuDomainGetPercpuStats): Drop pointless check for cpumap changes, and use correct number of cpus. --- Fixes the regression noticed here: https://www.redhat.com/archives/libvir-list/2012-October/msg01508.html src/qemu/qemu_driver.c | 26 +++----------------------- 1 file changed, 3 insertions(+), 23 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 18be7d9..f817319 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -13597,7 +13597,6 @@ qemuDomainGetPercpuStats(virDomainPtr domain, unsigned int ncpus) { virBitmapPtr map = NULL; - virBitmapPtr map2 = NULL; int rv = -1; int i, id, max_id; char *pos; @@ -13609,7 +13608,6 @@ qemuDomainGetPercpuStats(virDomainPtr domain, virTypedParameterPtr ent; int param_idx; unsigned long long cpu_time; - bool result; /* return the number of supported params */ if (nparams == 0 && ncpus != 0) @@ -13621,7 +13619,7 @@ qemuDomainGetPercpuStats(virDomainPtr domain, return rv; if (ncpus == 0) { /* returns max cpu ID */ - rv = max_id + 1; + rv = max_id; goto cleanup; } @@ -13648,11 +13646,7 @@ qemuDomainGetPercpuStats(virDomainPtr domain, id = start_cpu + ncpus - 1; for (i = 0; i <= id; i++) { - if (virBitmapGetBit(map, i, &result) < 0) - goto cleanup; - if (!result) { - cpu_time = 0; - } else if (virStrToLong_ull(pos, &pos, 10, &cpu_time) < 0) { + if (virStrToLong_ull(pos, &pos, 10, &cpu_time) < 0) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("cpuacct parse error")); goto cleanup; @@ -13680,22 +13674,9 @@ qemuDomainGetPercpuStats(virDomainPtr domain, if (getSumVcpuPercpuStats(group, priv->nvcpupids, sum_cpu_time, n) < 0) goto cleanup; - /* Check that the mapping of online cpus didn't change mid-parse. */ - map2 = nodeGetCPUBitmap(domain->conn, &max_id); - if (!map2 || !virBitmapEqual(map, map2)) { - virReportError(VIR_ERR_OPERATION_INVALID, "%s", - _("the set of online cpus changed while reading")); - goto cleanup; - } - sum_cpu_pos = sum_cpu_time; for (i = 0; i <= id; i++) { - if (virBitmapGetBit(map, i, &result) < 0) - goto cleanup; - if (!result) - cpu_time = 0; - else - cpu_time = *(sum_cpu_pos++); + cpu_time = *(sum_cpu_pos++); if (i < start_cpu) continue; if (virTypedParameterAssign(&params[(i - start_cpu) * nparams + @@ -13711,7 +13692,6 @@ cleanup: VIR_FREE(sum_cpu_time); VIR_FREE(buf); virBitmapFree(map); - virBitmapFree(map2); return rv; } -- 1.7.11.7

2 2

[libvirt] how to delete storage-pool entirely
by yue 26 Oct '12

26 Oct '12

it seems that if pools has the same path even with different uuid ,they will be recognize as one pool. if i have defined a pool with a special path, next time to define a new one(same path , different uuid) will fail. then i delete(rm -f ) pool.xml file, define the pool again , it fail too. so if there is cache for all pools been defined until next start libvirtd? my questions: 1. how to delete a pool entilely , include pool-xml file and possible cache (not to restart libvirtd).? 2.throught calling libvirt API , how to find the pool with a special path,no need to define it again? thanks

3 4

[libvirt] [PATCH] [trivial] documentation: HTML tag fix
by Philipp Hahn 26 Oct '12

26 Oct '12

Replace '%' by '&' for correct escaping of '>' in Domain specification. Signed-off-by: Philipp Hahn <hahn(a)univention.de> --- docs/formatdomain.html.in | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 2417943..c8da33d 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -3102,7 +3102,7 @@ qemu-kvm -net nic,model=? /dev/null provide their own way (outside of libvirt) to tag guest traffic onto specific vlans.) To allow for specification of multiple tags (in the case of vlan trunking), a - subelement, <code><tag%gt;</code>, specifies which vlan tag + subelement, <code><tag></code>, specifies which vlan tag to use (for example: <code><tag id='42'/></code>. If an interface has more than one <code><vlan></code> element defined, it is assumed that the user wants to do VLAN trunking -- 1.7.1

2 1

[libvirt] [PATCH v3 0/2] Qemu/Gluster support in Libvirt
by Harsh Prateek Bora 26 Oct '12

26 Oct '12

This patchset provides support for Gluster protocol based network disks. Changelog: v3: - RNG schema updated as required for unix transport [Paolo] - introduced another new attribute 'socket' for unix transport [Paolo] - Uses virURIFormat and virURIParse for URI parsing. [danpb] - updated documentation as required. [Jirka] v2: - Addressed review comments by Jiri - Updated patcheset as per new URI spec Ref: http://lists.gnu.org/archive/html/qemu-devel/2012-09/msg05199.html v1: - Initial prototype Harsh Prateek Bora (2): Qemu/Gluster: Add Gluster protocol as supported network disk formats. tests: Add tests for gluster protocol based network disks support docs/formatdomain.html.in | 24 +++- docs/schemas/domaincommon.rng | 35 ++++- src/conf/domain_conf.c | 72 ++++++++-- src/conf/domain_conf.h | 12 ++ src/libvirt_private.syms | 2 + src/qemu/qemu_command.c | 155 +++++++++++++++++++++ tests/qemuargv2xmltest.c | 1 + .../qemuxml2argv-disk-drive-network-gluster.args | 1 + .../qemuxml2argv-disk-drive-network-gluster.xml | 35 +++++ tests/qemuxml2argvtest.c | 2 + 10 files changed, 311 insertions(+), 28 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-network-gluster.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-network-gluster.xml -- 1.7.11.7

1 2

[libvirt] [PATCH] Revert "qemu: Do not require hostuuid in migration cookie"
by Jiri Denemark 26 Oct '12

26 Oct '12

This reverts commit 8d75e47edefdd77b86df1ee9af3cd5001d456f73. Libvirt was never released with support for migration cookies without hostuuid. --- src/qemu/qemu_migration.c | 31 +++++++++++++++---------------- 1 file changed, 15 insertions(+), 16 deletions(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index c15a75d..c4ac150 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -783,23 +783,22 @@ qemuMigrationCookieXMLParse(qemuMigrationCookiePtr mig, } if (!(tmp = virXPathString("string(./hostuuid[1])", ctxt))) { - VIR_WARN("Missing hostuuid element in migration data; cannot " - "detect migration to the same host"); - } else { - if (virUUIDParse(tmp, mig->remoteHostuuid) < 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, "%s", - _("malformed hostuuid element in migration data")); - goto error; - } - if (memcmp(mig->remoteHostuuid, mig->localHostuuid, - VIR_UUID_BUFLEN) == 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("Attempt to migrate guest to the same host %s"), - tmp); - goto error; - } - VIR_FREE(tmp); + virReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("missing hostuuid element in migration data")); + goto error; } + if (virUUIDParse(tmp, mig->remoteHostuuid) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("malformed hostuuid element in migration data")); + goto error; + } + if (memcmp(mig->remoteHostuuid, mig->localHostuuid, VIR_UUID_BUFLEN) == 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Attempt to migrate guest to the same host %s"), + tmp); + goto error; + } + VIR_FREE(tmp); /* Check to ensure all mandatory features from XML are also * present in 'flags' */ -- 1.7.12.4

2 1

[libvirt] [PATCH] qemu: Do not require hostuuid in migration cookie
by Jiri Denemark 26 Oct '12

26 Oct '12

Having hostuuid in migration cookie is a nice bonus since it provides an easy way of detecting migration to the same host. However, requiring it breaks backward compatibility with older libvirt releases. --- src/qemu/qemu_migration.c | 31 ++++++++++++++++--------------- 1 file changed, 16 insertions(+), 15 deletions(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index a2402ce..487182e 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -576,22 +576,23 @@ qemuMigrationCookieXMLParse(qemuMigrationCookiePtr mig, } if (!(tmp = virXPathString("string(./hostuuid[1])", ctxt))) { - virReportError(VIR_ERR_INTERNAL_ERROR, - "%s", _("missing hostuuid element in migration data")); - goto error; - } - if (virUUIDParse(tmp, mig->remoteHostuuid) < 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, - "%s", _("malformed hostuuid element in migration data")); - goto error; - } - if (memcmp(mig->remoteHostuuid, mig->localHostuuid, VIR_UUID_BUFLEN) == 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("Attempt to migrate guest to the same host %s"), - tmp); - goto error; + VIR_WARN("Missing hostuuid element in migration data; cannot " + "detect migration to the same host"); + } else { + if (virUUIDParse(tmp, mig->remoteHostuuid) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("malformed hostuuid element in migration data")); + goto error; + } + if (memcmp(mig->remoteHostuuid, mig->localHostuuid, + VIR_UUID_BUFLEN) == 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Attempt to migrate guest to the same host %s"), + tmp); + goto error; + } + VIR_FREE(tmp); } - VIR_FREE(tmp); /* Check to ensure all mandatory features from XML are also * present in 'flags' */ -- 1.7.12.4

4 5

[libvirt] [PATCH] maint: Sort .gitignore
by Peter Krempa 26 Oct '12

26 Oct '12

Sorting of the .gitignore file was broken after the last addition. After a clean build the scripts re-sort it making the working tree dirty. --- Pushed under trivial rule. --- .gitignore | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.gitignore b/.gitignore index 51d83a6..98ce398 100644 --- a/.gitignore +++ b/.gitignore @@ -1,4 +1,3 @@ -.#* *#*# *.#*# *.a @@ -18,6 +17,7 @@ *.rej *.s *~ +.#* .deps .gdb_history .git -- 1.7.12.4

1 0

[libvirt] [PATCH] virsh: Remove --flags from nodesuspend
by Jiri Denemark 26 Oct '12

26 Oct '12

We always expose individual bits from flags as separate options rather than exposing a raw flags options. Since virNodeSuspendForDuration does not currently support any flags, the only way of using this --flags options that would not fail is "--flags 0", which is equivalent to omitting the option. Thus it is highly unlikely anyone would actually be using it and removing it should be safe. --- tools/virsh-host.c | 10 +--------- tools/virsh.pod | 2 +- 2 files changed, 2 insertions(+), 10 deletions(-) diff --git a/tools/virsh-host.c b/tools/virsh-host.c index 0f9b3f3..2ea24ac 100644 --- a/tools/virsh-host.c +++ b/tools/virsh-host.c @@ -506,7 +506,6 @@ static const vshCmdOptDef opts_node_suspend[] = { {"target", VSH_OT_DATA, VSH_OFLAG_REQ, N_("mem(Suspend-to-RAM), " "disk(Suspend-to-Disk), hybrid(Hybrid-Suspend)")}, {"duration", VSH_OT_INT, VSH_OFLAG_REQ, N_("Suspend duration in seconds, at least 60")}, - {"flags", VSH_OT_INT, VSH_OFLAG_NONE, N_("Suspend flags, 0 for default")}, {NULL, 0, 0, NULL} }; @@ -516,7 +515,6 @@ cmdNodeSuspend(vshControl *ctl, const vshCmd *cmd) const char *target = NULL; unsigned int suspendTarget; long long duration; - unsigned int flags = 0; if (vshCommandOptString(cmd, "target", &target) < 0) { vshError(ctl, _("Invalid target argument")); @@ -528,11 +526,6 @@ cmdNodeSuspend(vshControl *ctl, const vshCmd *cmd) return false; } - if (vshCommandOptUInt(cmd, "flags", &flags) < 0) { - vshError(ctl, _("Invalid flags argument")); - return false; - } - if (STREQ(target, "mem")) suspendTarget = VIR_NODE_SUSPEND_TARGET_MEM; else if (STREQ(target, "disk")) @@ -549,8 +542,7 @@ cmdNodeSuspend(vshControl *ctl, const vshCmd *cmd) return false; } - if (virNodeSuspendForDuration(ctl->conn, suspendTarget, duration, - flags) < 0) { + if (virNodeSuspendForDuration(ctl->conn, suspendTarget, duration, 0) < 0) { vshError(ctl, "%s", _("The host was not suspended")); return false; } diff --git a/tools/virsh.pod b/tools/virsh.pod index 76f32c2..321dbe3 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -294,7 +294,7 @@ statistics during 1 second. Returns memory stats of the node. If I<cell> is specified, this will prints specified cell statistics only. -=item B<nodesuspend> [I<target>] [I<duration>] [I<flags>] +=item B<nodesuspend> [I<target>] [I<duration>] Puts the node (host machine) into a system-wide sleep state such as Suspend-to-RAM, Suspend-to-Disk or Hybrid-Suspend and sets up a -- 1.7.12.4

6 6

[libvirt] [PATCH] maint: ignore unsaved emacs files
by Eric Blake 26 Oct '12

26 Oct '12

I did a 'git add .', then realized that it ended up trying to add the emacs lock file for a corresponding file that I had not yet saved all my edits; thankfully I noticed it in time. Since we already exclude other temporary files, this makes the most sense for preventing such a mistake from actually hitting upstream. * .gitignore: Add .#* to the exclude list. --- If you don't like this patch, then I will be updating my local .git/info/exclude file; I could understand someone wanting to reject this patch on the grounds that 'make' does not create any temporary files under this name, but only my choice of editor (and not everyone shares that choice in editor). So I'll wait for an ack instead of claiming the trivial rule. .gitignore | 1 + 1 file changed, 1 insertion(+) diff --git a/.gitignore b/.gitignore index 804eda4..51d83a6 100644 --- a/.gitignore +++ b/.gitignore @@ -1,3 +1,4 @@ +.#* *#*# *.#*# *.a -- 1.7.11.7

3 5

[libvirt] [PATCHv3] virNodeGetCPUMap: Add python binding
by Viktor Mihajlovski 26 Oct '12

26 Oct '12

Added a method getCPUMap to virConnect. It can be used as follows: import libvirt import sys import os conn = libvirt.openReadOnly(None) if conn == None: print 'Failed to open connection to the hypervisor' sys.exit(1) try: (cpus, cpumap, online) = conn.getCPUMap(0) except: print 'Failed to extract the node cpu map information' sys.exit(1) print 'CPUs total %d, online %d' % (cpus, online) print 'CPU map %s' % str(cpumap) del conn print "OK" sys.exit(0) Signed-off-by: Viktor Mihajlovski <mihajlov(a)linux.vnet.ibm.com> --- Rework of patch 9/9 of initial virNodeGetCPUMap series. squashed in Eric Blake's changes added missing error handling reordered return value processing Note: The code is a bit verbose, e.g. the NULL checks, but that seems to be the prevailing style here. --- python/libvirt-override-api.xml | 6 +++ python/libvirt-override.c | 70 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 76 insertions(+), 0 deletions(-) diff --git a/python/libvirt-override-api.xml b/python/libvirt-override-api.xml index b76fb4e..e54701c 100644 --- a/python/libvirt-override-api.xml +++ b/python/libvirt-override-api.xml @@ -542,5 +542,11 @@ <arg name='conn' type='virConnectPtr' info='pointer to the hypervisor connection'/> <arg name='flags' type='int' info='unused, always pass 0'/> </function> + <function name='virNodeGetCPUMap' file='python'> + <info>Get node CPU information</info> + <return type='str *' info='(cpunum, online, cpumap) on success, None on error'/> + <arg name='conn' type='virConnectPtr' info='pointer to the hypervisor connection'/> + <arg name='flags' type='int' info='unused, pass 0'/> + </function> </symbols> </api> diff --git a/python/libvirt-override.c b/python/libvirt-override.c index 83dc925..0609803 100644 --- a/python/libvirt-override.c +++ b/python/libvirt-override.c @@ -6397,6 +6397,75 @@ cleanup: return ret; } +static PyObject * +libvirt_virNodeGetCPUMap(PyObject *self ATTRIBUTE_UNUSED, + PyObject *args) +{ + virConnectPtr conn; + PyObject *pyobj_conn; + PyObject *ret = NULL; + PyObject *pycpumap = NULL; + PyObject *pyused = NULL; + PyObject *pycpunum = NULL; + PyObject *pyonline = NULL; + int i_retval; + unsigned char *cpumap = NULL; + unsigned int online = 0; + unsigned int flags; + int i; + + if (!PyArg_ParseTuple(args, (char *)"Oi:virNodeGetCPUMap", + &pyobj_conn, &flags)) + return NULL; + conn = (virConnectPtr) PyvirConnect_Get(pyobj_conn); + + LIBVIRT_BEGIN_ALLOW_THREADS; + i_retval = virNodeGetCPUMap(conn, &cpumap, &online, flags); + LIBVIRT_END_ALLOW_THREADS; + + if (i_retval < 0) + return VIR_PY_NONE; + + if ((ret = PyTuple_New(3)) == NULL) + goto error; + + /* 0: number of CPUs */ + if ((pycpunum = PyLong_FromLong(i_retval)) == NULL || + PyTuple_SetItem(ret, 0, pycpunum) < 0) + goto error; + + /* 1: CPU map */ + if ((pycpumap = PyList_New(i_retval)) == NULL) + goto error; + + for (i = 0; i < i_retval; i++) { + if ((pyused = PyBool_FromLong(VIR_CPU_USED(cpumap, i))) == NULL) + goto error; + if (PyList_SetItem(pycpumap, i, pyused) < 0) + goto error; + } + + if (PyTuple_SetItem(ret, 1, pycpumap) < 0) + goto error; + + /* 2: number of online CPUs */ + if((pyonline = PyLong_FromLong(online)) == NULL || + PyTuple_SetItem(ret, 2, pyonline) < 0) + goto error; + +cleanup: + VIR_FREE(cpumap); + return ret; +error: + Py_XDECREF(ret); + Py_XDECREF(pycpumap); + Py_XDECREF(pyused); + Py_XDECREF(pycpunum); + Py_XDECREF(pyonline); + ret = NULL; + goto cleanup; +} + /************************************************************************ * * @@ -6514,6 +6583,7 @@ static PyMethodDef libvirtMethods[] = { {(char *) "virDomainGetDiskErrors", libvirt_virDomainGetDiskErrors, METH_VARARGS, NULL}, {(char *) "virNodeGetMemoryParameters", libvirt_virNodeGetMemoryParameters, METH_VARARGS, NULL}, {(char *) "virNodeSetMemoryParameters", libvirt_virNodeSetMemoryParameters, METH_VARARGS, NULL}, + {(char *) "virNodeGetCPUMap", libvirt_virNodeGetCPUMap, METH_VARARGS, NULL}, {NULL, NULL, 0, NULL} }; -- 1.7.0.4

2 2