QEMU supports a bunch of CPUID features that are tied to the kvm CPUID
nodes rather than the processor's.  They are "kvmclock",
"kvm_nopiodelay", "kvm_mmu", "kvm_asyncpf".  These are not known to
libvirt and their CPUID leaf might move if (for example) the Hyper-V
extensions are enabled. Hence their handling would anyway require some
special-casing.

However, among these the most useful is kvmclock; an additional
"property" of this feature is that a <timer> element is a better model
than a CPUID feature.  Although, creating part of the -cpu command-line
from something other than the <cpu> XML element introduces some
ugliness.
Reviewed-by: NJiri Denemark <jdenemar@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Add kvmclock timer to documentation, schema and parsers.  Keep the
platform timer first since it is kind of special, and alphabetize
the others when possible (i.e. when it does not change the ABI).
Reviewed-by: NJiri Denemark <jdenemar@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Avoid creating an empty <cpu> element when the QEMU command-line simply
specifies the default "-cpu qemu32" or "-cpu qemu64".

This requires the previous patch, which lets us represent "-cpu qemu32"
as <os arch='i686'> in the generated XML.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The qemu32 CPU model is chosen based on the <os arch=...> name when
creating the QEMU command line for a 64-bit host.  For the opposite
transformation we can test the guest CPU model for the "lm" feature.
If it is absent, def->os.arch needs to be corrected.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

When running under KVM, the arch is usually set to i686 because
the name of the emulator is not qemu-system-x86_64.  Use the host
arch instead.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Recently (or not so recently) QEMU added the kvm32 and kvm64
architectures, representing a least common denominator of all
hosts that can run KVM.  Add them to the machine map.

Also, some features that TCG supports were added to qemu64.
Add them to the cpu_map.xml whenever KVM is guaranteed to support
those.  We still have to leave some out, because they would not
be available to guests running on older hosts.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The qemu developers have made it clear that modern qemu will no
longer guarantee human monitor command stability; furthermore,
some features, such as async events, are only supported via qmp.
If we are compiled without support for handling JSON, we cannot
expect to sanely interact with modern qemu.

However, things must continue to build on RHEL 5, where qemu
is stuck at 0.10, and where yajl is not available.

Another benefit of this patch: future additions of new monitor
commands need only focus on qemu_monitor_json.c, instead of
also wasting time with qemu_monitor_text.c.

* src/qemu/qemu_capabilities.c (qemuCapsComputeCmdFlags): Report
error if yajl is missing but qemu requires qmp.
(qemuCapsParseHelpStr): Propagate error.
(qemuCapsExtractVersionInfo): Update caller.
* tests/qemuhelptest.c (testHelpStrParsing): Likewise.

I'm getting tired of remembering to backport RHEL-specific
patches when building upstream libvirt on RHEL 6.x or CentOS.
All the affected versions of RHEL qemu-kvm have backported
enough patches to a) make JSON useful, and b) modify the
-help text to mention libvirt as the preferred interface;
which means this string in the help output is a reliable
indicator that we can outsmart a strict version check,
even when upstream qemu 0.12 lacked the needed features.

* src/qemu/qemu_capabilities.c (qemuCapsComputeCmdFlags):
Recognize particular help string present when enough features were
backported to be worth using JSON.
* tests/qemuhelptest.c (mymain): Update tests accordingly.

Compare two filters' XML for equality and only rebuild/instantiate the new
filter if the new and current filters are found to be different. This
improves performance during an update of a filter with no obvious change
or the reloading of filters during a 'kill -SIGHUP'

Introduce a function that rebuilds all running VMs' filters. Call
this function when reloading the nwfilter driver.

This addresses a problem introduced by the 2nd patch that typically
causes no filters to be reinstantiate anymore upon driver reload
since their XML has not changed. Yet the current behavior is that
upon a SIGHUP all filters get reinstantiated.

QEMU always sends details about all available block devices as an answer
for "info block"/"query-block" command. On the other hand, our
qemuMonitorGetBlockInfo was made for a single block devices queries
only. Thus, when asking for multiple devices, we asked qemu multiple
times to always get the same answer from which different parts were
filtered. This patch makes qemuMonitorGetBlockInfo return a hash table
of all block devices, which may later be used for getting details about
specific devices.

Without this, virt-aa-helper would segfault in -c or -r commands.

Added a new field "vm-pid" to the VIRT_CONTROL audit record. This information
is useful to correlated another audit events to the events generated by
libvirt.

Convert daemon code to handle 64-bit pid_t (even though at the
moment, it is not compiled on mingw).

* daemon/remote.c (remoteDispatchAuthList)
(remoteDispatchAuthPolkit): Print pid_t via %lld.

Using snprintf to build up argv seems archaic.

* daemon/remote.c (remoteDispatchAuthPolkit): Modernize command call.

On RHEL5, I got:
util/virrandom.c:66: warning: nested extern declaration of '_gl_verify_function66' [-Wnested-externs]

The fix is to hoist the verify earlier.  Also some other hodge-podge
fixes I noticed while reviewing Dan's recent series.

* .gitignore: Ignore new test.
* src/util/cgroup.c: Bump copyright year.
* src/util/virhash.c: Fix typo in description.
* src/util/virrandom.c (virRandomBits): Mark doc comment, and
hoist assert to silence older gcc.

Some files are using uint32_t or int64_t without including
stdint.h which defines them. Fix this.

* src/libvirt_public.syms: Death to tabs
* cfg.mk: Check .syms files for tabs

Recent discussions have illustrated the potential for DOS attacks
with the hash table implementations used by most languages and
libraries.

   https://lwn.net/Articles/474912/

libvirt has an internal hash table impl, and uses hash tables for
a variety of purposes. The hash key generation code is pretty
simple and thus not strongly collision resistant.

This patch replaces the current libvirt hash key generator with
the (public domain) Murmurhash3 code. In addition every hash
table now gets a random seed value which is used to perturb the
hashing code. This should make it impossible to mount any
practical attack against libvirt hashing code.

* bootstrap.conf: Import bitrotate module
* src/Makefile.am: Add virhashcode.[ch]
* src/util/util.c: Make virRandom() return a fixed 32 bit
  integer value.
* src/util/hash.c, src/util/hash.h, src/util/cgroup.c: Replace
  hash code generation with a call to virHashCodeGen()
* src/util/virhashcode.h, src/util/virhashcode.c: Add a new
  virHashCodeGen() API using the Murmurhash3 algorithm.

In preparation for the patch to include Murmurhash3, which
introduces a virhashcode.h and virhashcode.c files, rename
the existing hash.h and hash.c to virhash.h and virhash.c
respectively.

In preparation for conversion over to use the Murmurhash3
algorithm, convert various virHash APIs to use size_t or
uint32 for their return values/parameters, instead of the
variable size 'unsigned long' or 'int' types

The old virRandom() API was not generating good random numbers.
Replace it with a new API virRandomBits which instead of being
told the upper limit, gets told the number of bits of randomness
required.

* src/util/virrandom.c, src/util/virrandom.h: Add virRandomBits,
  and move virRandomInitialize
* src/util/util.h, src/util/util.c: Delete virRandom and
  virRandomInitialize
* src/libvirt.c, src/security/security_selinux.c,
  src/test/test_driver.c, src/util/iohelper.c: Update for
  changes from virRandom to virRandomBits
* src/storage/storage_backend_iscsi.c: Remove bogus call
  to virRandomInitialize & convert to virRandomBits

The domain schema enforced restrictions on the domain name string that
the code doesn't. This patch relaxes the check, leaving the restrictions
on the driver or hypervisor. The only invalid character is a newline.

Currently, we support only filling a volume with zeroes on wiping.
However, it is not enough as data might still be readable by
experienced and equipped attacker. Many technical papers have been
written, therefore we should support other wiping algorithms.

Typo introduced in commit 4e9953a4, and remained in 6fba577e.

* tools/virsh.pod (snapshot-create): Fix pod error.

Fix a few warnings with mingw64 x86_64.

In file included from ../gnulib/lib/unistd.h:51:0,
                 from ../src/util/util.h:30,
                 from rpc/virkeepalive.c:29:
/usr/x86_64-w64-mingw32/sys-root/mingw/include/winsock2.h:15:2: warning: #warning Please include winsock2.h before windows.h [-Wcpp]

Reported by Marc-André Lureau.

* src/util/threads-win32.h (includes): Pick up winsock2.h before
windows.h, as required by mingw64.

Fixes the following warning:
util/virterror.c:1242:31: warning: declaration of 'errcode' shadows a global declaration [-Wshadow]

Add missing function if !HAVE_GETPWUID_R.

On F16 at least, empty volume groups don't have a directory under /dev.
The directory only appears once a logical volume is created.

This tickles some behavior in BackendStablePath which ends with
libvirt sleeping for 5 seconds while waiting for the directory to appear.
This causes all sorts of problems for the virStorageVolLookupByPath API
which virtinst uses, even if trying to resolve a path that is independent
of the logical pool.

In reality we don't even need to do that checking since logical pools
always have a stable target path. Short circuit the polling in that
case.

Fixes bug 782261

Systemd detects containers based on whether they have
an environment variable starting with 'container=lxc';
using a longer name fits the expectations, while also
allowing detection of who created the container.

Requested by Lennart Poettering, in response to
https://bugs.freedesktop.org/show_bug.cgi?id=45175

* src/lxc/lxc_container.c (lxcContainerBuildInitCmd): Add another
env-var.

The current setup code for LXC is bind mounting /dev/pts/ptmx
on top of a character device /dev/ptmx. This is denied by SELinux
policy and is just wrong. The target of a bind mount should just
be a plain file

* src/lxc/lxc_container.c: Don't bind /dev/pts/ptmx onto
  a char device

Add a virFileTouch API which ensures that a file will always
exist, even if zero length

* src/util/virfile.c, src/util/virfile.h,
  src/libvirt_private.syms: Introduce virFileTouch

to cmdSnapshotCreate and cmdSnapshotCreateAs.

With this flag, virDomainSnapshotCreate will use fs-freeze and
fs-thaw guest agent commands to quiesce guest's disks.

These functions simply issue command to guest agent which
should freeze or unfreeze all file systems within guest.

Direct boot (using kernel, initrd, and command line) is used by
virt-install/virt-manager for network install. While any bootindex has
no direct effect since -kernel is always first, we need it as a hint for
SeaBIOS to present disks in the same order as they will be presented
during normal boot.