For virtio disks and interfaces, qemu allows users to enable or disable
ioeventfd feature. This means, qemu can execute domain code, while
another thread waits for I/O event. Basically, in some cases it is win,
in some loss. This feature is available via 'ioeventfd' attribute in disk
and interface <driver> element. It accepts 'on' and 'off'. Leaving this
attribute out defaults to hypervisor decision.

When building rpms for newer Fedora or RHEL, take advantage of the
newer netcf packaging to guarantee interface snapshot support.

* libvirt.spec.in (BuildRequires): Bump minimum version on
platforms that support netcf 0.1.8.

https://bugzilla.redhat.com/show_bug.cgi?id=682121

Gettext reserves the empty string for internal use, and it must
not be passed through _().  We were violating this for commands
that (for whatever reason) used "" for their description.

* tools/virsh.c (vshCmddefHelp): Don't translate empty string.
Reported by Tatsuo Kawasaki.

The following patch addresses the problem that when a PASSTHROUGH
mode DIRECT NIC connection is made the MAC address of the NIC is
not automatically set and reset to the configured VM MAC and
back again.

The attached patch fixes this problem by setting and resetting the MAC
while remembering the previous setting while the VM is running.
This also works if libvirtd is restarted while the VM is running.

the patch passes make syntax-check

These functions aren't intended to be called directly by users, so mark
them as private.

While we're at it, remove unneeded exception handling, and break some
long lines.

If registering our own event loop implementation written in python,
any handles or timeouts callbacks registered by libvirt C code must
be wrapped in a python function. There is some argument trickery that
makes this all work, by wrapping the user passed opaque value in
a tuple, along with the callback function.

Problem is, the current setup requires the user's event loop to know
about this trickery, rather than just treating the opaque value
as truly opaque.

Fix this in a backwards compatible manner, and adjust the example
python event loop to do things the proper way.

Since we virEventRegisterDefaultImpl is now a public API, callers need
a way to invoke the default registered Handle and Timeout functions. We
already have general functions for these internally, so promote
them to the public API.

v2:
    Actually add APIs to libvirt.h

Pure python implementation. The handler callbacks have been altered
a bit compared to the C API: RecvAll doesn't pass length of the data read
since that can be trivially obtained from python string objects, and SendAll
requires the handler to return the string data to send rather than
store the data in a string pointer.

The return values for the python version are different that the C version
of virStreamSend: on success we return a string, an error raises an exception,
and if the stream would block we return int(-2). We need to do this
since strings aren't passed by reference in python.

* virDomainDefParse: There is a goto label "no_memory", which
reports OOM error, and then fallthrough label "error". This
patch changes things like following:

    virReportOOMError();
    goto error;

into:

    goto no_memory;

Free def->numatune.memory.nodemask in virDomainDefFree.

v2:
    Don't generate virStreamFree

In a couple instances we have to mark a debug variable as ATTRIBUTE_UNUSED
to avoid warnings.

v2:
    Use #if 0 to comment out debug define

Similar to pool-create-as.

* tools/virsh.c (cmdSnapshotCreateAs): Add --print-xml.
* tools/virsh.pod: Document it.

Producing an xml file just for name and description fields is
overkill; this makes life easier from virsh.

* tools/virsh.c (cmdSnapshotCreateAs): New command.
(snapshotCmds): Install it.
* tools/virsh.pod: Document it.

* tools/virsh.c (info_snapshot_create, info_save): Clarify
description.
* tools/virsh.pod (save): Likewise.

The 'char *cur' variable was being assigned from a
'const char *' string, thus discarding constness.
As well as causing a compile warning, it masked a
piece of code which attempts to assign to the
previously const string.

* tools/virsh.c: Fix const-ness of 'cur' variable in vcpupin

Removes special case code from the generator and handle additional
methods.

The generated version of remoteDispatchDomainPinVcpu(Flags) has no
length check, but this check was useless anyway as it was applied to
data that was already deserialized from its XDR form.

It was already generatable but skipped.

Pinning to all physical cpus means resetting, hence it is preferable to
delete vcpupin setting of XML.

This patch changes qemu driver to delete vcpupin setting by invoking
virDomainVcpupinDel API when pinning the specified virtual cpu to
all host physical cpus.
Signed-off-by: NTaku Izumi <izumi.taku@jp.fujitsu.com>

This patch add the private API (virDomainVcpupinDel).
This API can delete the vcpupin setting of a specified virtual cpu.
Signed-off-by: NTaku Izumi <izumi.taku@jp.fujitsu.com>

When resetting vcpupin setting, we have to specify all host physical
cpus as a cpulist parameter of virsh vcpupin command. It's a little
tedious.

This patch changes to allow to receive the special keyword 'r' as a cpulist
parameter of virsh vcpupin command when resetting vcpupin setting.

If you set the following:

 # virsh vcpupin VM 0 r

the vcpu0 will be pinned to all physical cpus.
Signed-off-by: NTaku Izumi <izumi.taku@jp.fujitsu.com>

When using vcpupin command, we have to speficy comma-separated list as cpulist,
but this is tedious in case the number of phsycal cpus is large.
This patch improves this by introducing special markup "-" and "^" which are
similar to XML schema of "cpuset" attribute.

The example:

 # virsh vcpupin Guest 0 0-15,^8

 is identical to

 # virsh vcpupin Guest 0 0,1,2,3,4,5,6,7,9,10,11,12,13,14,15

NOTE: The expression is sequentially evaluated, so "0-15,^8" is not identical
to "^8,0-15".
Signed-off-by: NTaku Izumi <izumi.taku@jp.fujitsu.com>

The XenAPI driver works like the ESX and PHyp driver by using its
own HTTPS based remote protocol.

Only add tests for qemuxmlargvtest.c, as there is no qemu command line
for numatune XML, just want to make sure the XML could be validated
well.

Implemented as setting NUMA policy between fork and exec as a hook,
using libnuma. Only support memory tuning on domain process currently.

For the nodemask out of range, will report soft warning instead of
hard error in libvirt layer. (Kernel will be silent as long as one
of set bit in the nodemask is valid on the host. E.g. For a host
has two NUMA nodes, kernel will be silent for nodemask "01010101").
So, soft warning is the only thing libvirt can do, as one might want
to specify the numa policy prior to a node that doesn't exist yet,
however, it may come as hotplug soon.

* src/conf/domain_conf.h: Introduce one new struct for representing
NUMA tuning related stuffs.

* src/conf/domain_conf.c: Parse and format numatune XML.

Example XML:

<numatune>
  <memory mode="strict" nodeset="0-10,^4"/>
</numatune>

Please enter the commit message for your changes. Lines starting

Add --config, --live and --current for command blkiotune

When building libvirt without libvirtd, we will receive the following error
message:

make[3]: Entering directory `/home/wency/rpmbuild/BUILD/libvirt-0.9.2/tools'
  CC     virsh-virsh.o
  CC     virsh-console.o
  GEN    virt-xml-validate
  GEN    virt-pki-validate
  CCLD   virsh
./src/.libs/libvirt.so: undefined reference to `numa_available'
./src/.libs/libvirt.so: undefined reference to `numa_max_node'
collect2: ld returned 1 exit status

The reason is that: we check numactl only when building qemu driver, and qemu
driver will not be built when bulding without libvirtd. So with_numactl's
value is check and we will not link libnuma.so.

In the other function, we call numa_available() and numa_max_node() only
when HAVE_NUMACTL is 1. We should do the same check in the function nodeGetMemoryStats().

During a savevm operation, libvirt will now use fd migration if qemu
supports it. When the AppArmor driver is enabled, AppArmorSetFDLabel()
is used but since this function simply returns '0', the dynamic AppArmor
profile is not updated and AppArmor blocks access to the save file. This
patch implements AppArmorSetFDLabel() to get the pathname of the file by
resolving the fd symlink in /proc, and then gives that pathname to
reload_profile(), which fixes 'virsh save' when AppArmor is enabled.

Reference: https://launchpad.net/bugs/795800

The domain lock manager forgot to include a bunch of checks
for NULL which could occur on OOM

* src/locking/domain_lock.c: Add checks for NULL