The -mem-prealloc flag should be used when using large pages
This ensures qemu tries to allocate all required memory immediately,
rather than when first used. The latter mode will crash qemu
if hugepages aren't available when accessed, while the former
should gracefully fallback to non-hugepages.

* src/qemu/qemu_conf.c: add -mem-prealloc flag to qemu command line
  when using large pages

xen-unstable c/s 20685 changed the domctl interface, adding a field to
xen_domctl_getdomaininfo structure.  This additional field causes stack
corruption in libvirt.  xen-unstable c/s 20711 rightly bumped the domctl
interface version so it is at least possible to handle the new field.
This change accounts for shr_pages field added to xen_domctl_getdomaininfo
structure.

* src/network/bridge_driver.c (networkBuildDnsmasqArgv): Correct
test for NULL *argv.

The MAC addresses with 00:50:56 prefix are split into several ranges:

  00:50:56:00:00:00 - 00:50:56:3f:ff:ff  'static' range (manually assigned)
  00:50:56:80:00:00 - 00:50:56:bf:ff:ff  'vpx' range (assigned by a VI Client)

Erroneously the 'vpx' range was assumed to be larger and to occupy the
remaining addresses of the 00:50:56 prefix that are not part of the 'static'
range.

00:50:56 was used as prefix for generated MAC addresses, this is not possible
anymore, because there are gaps in the allowed ranges. Therefore, change the
prefix to 00:0c:29 which is the prefix for auto generated MAC addresses anyway.

Allow arbitrary MAC addresses to be used and set the checkMACAddress VMX option
to false in case the MAC address doesn't fall into any predefined range.

* docs/drvesx.html.in: update website accordingly
* src/esx/esx_driver.c: set the auto generation prefix to 00:0c:29
* src/esx/esx_vmx.c: fix MAC address range handling and allow arbitrary MAC
  addresses
* tests/vmx2xml*, tests/xml2vmx*: add some basic MAC address range tests

The data passed to the callback is not guaranteed to be zero terminated,
take care of that by coping the data and adding a zero terminator.

Also dump the data for other types than CURLINFO_TEXT.

Set CURLOPT_VERBOSE to 1 so the debug callback is called when enabled.

A domain with virtualHW version 4 is allowed on an ESX 4.0 server.
If a domain is migrated from an ESX 3.5 server to an ESX 4.0 server
then the virtualHW version stays the same. So a ESX 4.0 server can
host domains with virtualHW version 4.

This invalid free results in heap corruption. Some symptoms I saw
because of this were libvirtd crashing and virt-manager hanging
while trying to enumerate devices.

* src/xen/xend_internal.c (xenDaemonOpen_unix): Do not dereference
a NULL "conn".  Move first deref to follow the "conn == NULL" test.

The behavior for the qemu balloon device has changed.  Formerly, a virtio
balloon device was provided by default.  Now, '-balloon virtio' must be
specified on the command line to enable it.  This patch causes libvirt to
add '-balloon virtio' to the command line whenever the -balloon option is
available.
* src/qemu/qemu_conf.c src/qemu/qemu_conf.h: check for the new flag and
  add "-baloon vitio" to qemu command when needed
* tests/qemuhelptest.c: add the new flag for detection

This patch removes the call to vol update after the volume build completes.
The update call is currently meaningless anyway because the vol build is passed
a copy of the definition, so the update result is thrown away.  More
importantly, if the user specified a selinux label for the volume, the update
call results in a double free of the label
* src/storage/storage_backend_fs.c: remove the update call

This change makes the 'info chardev' parser ignore any trailing
whitespace on a line. This fixes a specific problem handling a '\r\n'
line ending.

* src/qemu/qemu_monitor_text.c: Ignore trailing whitespace in
  'info chardev' output.

* src/vbox/vbox_tmpl.c (vboxDomainCreateXML): "Unref" the domain
upon failure.  Patch by Matthias Bolte.

* src/vbox/vbox_tmpl.c (vboxDomainCreateXML): Don't call
vboxDomainUndefine on a NULL "dom".

* src/qemu/qemu_driver.c (qemudDomainMigratePrepare2): Remove useless
test of always-non-NULL uri_out parameter.  Use ATTRIBUTE_NONNULL to
inform tools.

* src/qemu/qemu_driver.c (qemudDomainAttachDevice): Call
virCgroupDenyDevicePath only if cgroup is non-NULL.

* src/openvz/openvz_conf.c (openvzLoadDomains): Avoid NULL deref
of "dom".

All other stateful drivers are linked directly to libvirtd
instead of libvirt.so. Link the secret driver to libvirtd too.

* daemon/Makefile.am: link the secret driver to libvirtd
* daemon/libvirtd.c: add #ifdef WITH_SECRETS blocks
* src/Makefile.am: don't link the secret driver to libvirt.so
* src/libvirt_private.syms: remove the secretRegister symbol

MinGW's ld gives an error when trying to export undefined symbols.

Alter the offending variable names to fix this.

* src/Makefile.am src/cpu/cpu_map.c: rename in Makefile and in the
  cpu map loading code

If using a remote access, sometimes an RPC entry point is not
available, and currently we just end up with a raw:
    error: unknown procedure: xxx
error, while this should be more cleanly reported as an unsupported
entry point like for local access

* src/remote/remote_driver.c: convert missing remote entry points into
  the unsupported feature error

When querying about a domain from 0.3.3 (or RHEL 5.3) domain located
on a 0.6.3 (RHEL-5) machine, the errors are not properly reported.
This patch from Olivier Fourdan <ofourdan@redhat.com> , slightly
modified to not change the semantic when the domain os details cannot
be provided

* src/xen/proxy_internal.c src/xen/xen_hypervisor.c: add some missing
  error reports

Found while trying to cross-compile libvirt on Fedora 12 for Windows.
gnulib redefines 'close' to 'close_used_without_including_unistd_h'
in sys/socket.h if winsock2.h is present and unistd.h has not been
included before sys/socket.h. Reorder some includes to fix this.

ATTRIBUTE_SENTINEL was defined for GCC >= 4.0 only, resulting in
compile errors when using GCC < 4.0.

Also backup and restore the original ctxt->node value in x86FeatureLoad.

* src/qemu/qemu_conf.h: Remove QEMU_CMD_FLAG_0_12 and just leave
  the lone JSON flag
* src/qemu/qemu_conf.c: Enable JSON on QEMU 0.13 or later, but
  leave it disabled for now

* src/qemu/qemu_conf.c: CPU models should be probed only if the
  a guest's XML contains CPU model, not each time a qemu command line
  is generated.

* src/Makefile.am: install it in $(pkgdatadir) i.e. /usr/share/libvirt/
  usually

The Xen code for making HVM VT-d PCI passthrough attach and detach
wasn't working properly:

1)  In xenDaemonAttachDevice(), we were always trying to reconfigure
a PCI passthrough device, even the first time we added it.  This was
because the code in virDomainXMLDevID() was not checking xenstore for
the existence of the device, and always returning 0 (meaning that
the device already existed).

2)  In xenDaemonDetachDevice(), we were trying to use "device_destroy"
to detach a PCI device.  While you would think that is the right
method to call, it's actually wrong for PCI devices.  In particular,
in upstream Xen (and soon in RHEL-5 Xen), device_configure is actually
used to destroy a PCI device.

To fix the attach
problem I add a lookup into xenstore to see if the device we are
trying to attach already exists.  To fix the detach problem I change
it so that for PCI detach (only), we use device_configure with the
appropriate sxpr to do the detachment.

* src/xen/xend_internal.c: don't use device_destroy for PCI devices
  and fix the other issues.
* src/xen/xs_internal.c src/xen/xs_internal.h: add
  xenStoreDomainGetPCIID()

The XML XPath for detecting JSON in the running VM statefile was
wrong causing all VMs to get JSON mode enabled at libvirtd restart.

In addition if a VM was running a JSON enabled QEMU once, and then
altered to point to a non-JSON enabled QEMU later the 'monJSON'
flag would not get reset to 0.

* src/qemu/qemu_driver.c: Fix setting/detection of JSON mode

The code for connecting to a server tries each socket in turn
until it finds one that connects. Unfortunately for TLS sockets
if it connected, but failed TLS handshake it would treat that
as a failure to connect, and try the next socket. This is bad,
it should have reported the TLS failure immediately.

$ virsh -c qemu://somehost.com/system
error: unable to connect to libvirtd at 'somehost.com': Invalid argument
error: failed to connect to the hypervisor

$ ./tools/virsh -c qemu://somehost.com/system
error: server certificate failed validation: The certificate hasn't got a known issuer.
error: failed to connect to the hypervisor

 * src/remote/remote_driver.c: Stop trying to connect if the
   TLS handshake fails

* src/qemu/qemu_driver.c: The 'if' statement is supposed to check return
  value of pci = pciGetDevice(conn, ...); , but check uses if (!dev),
  fix it

Use a dynamically sized xdr_array to pass memory stats on the wire.  This
supports the addition of future memory stats and reduces the message size
since only supported statistics are returned.

* src/remote/remote_protocol.x: provide defines for the new entry point
* src/remote/remote_driver.c daemon/remote.c: implement the client and
  server side
* daemon/remote_dispatch_args.h daemon/remote_dispatch_prototypes.h
  daemon/remote_dispatch_ret.h daemon/remote_dispatch_table.h
  src/remote/remote_protocol.c src/remote/remote_protocol.h: generated
  stubs

Support for memory statistics reporting is accepted for qemu inclusion.
Statistics are reported via the monitor command 'info balloon' as a comma
seprated list:

(qemu) info balloon
balloon: actual=1024,mem_swapped_in=0,mem_swapped_out=0,major_page_faults=88,minor_page_faults=105535,free_mem=1017065472,total_mem=1045229568

Libvirt, qemu, and the guest operating system may support a subset of the
statistics defined by the virtio spec.  Thus, only statistics recognized by
components will be reported.

* src/qemu/qemu_driver.c src/qemu/qemu_monitor_text.[ch]: implement the
  new entry point by using info balloon monitor command

* src/libvirt.c: implement the main entry point
* src/libvirt_public.syms: add it to the exported symbols

Set up the types for the domainMemoryStats function and insert it into the
virDriver structure definition.  Because of static initializers, update
every driver and set the new field to NULL.

* include/libvirt/libvirt.h.in: new API
* src/driver.h src/*/*_driver.c src/vbox/vbox_tmpl.c: add the new
  entry to the driver structure
* python/generator.py: fix compiler errors, the actual python binding is
  implemented later