Perform all the appropriate plumbing.

When qemu/KVM VMs are paused manually through a monitor not-owned by libvirt,
libvirt will think of them as "paused" event after they are resumed and
effectively running. With this patch the discrepancy goes away.

This is meant to address bug 892791.
Signed-off-by: NAndres Lagar-Cavilla <andres@lagarcavilla.org>

Libvirt's HTML documentation is not as easy to the eyes as it could
be since long text has no visual breaks.

Take advantage of the formatting in documentation comments and wrap
each part separated by two consecutive \n into a HTML <p> element.

Coverity determined that 'emulator' could no longer be set and determined the
code was dead.  Looking through the history, I discovered commit-id ed769e18
removed code originally added by commit-id 9237e955 and further modified by
commit-id 6a7e7c4f.

Commit id f8ab364c removed ability to run this driver unprivileged. Coverity
detected the check and flagged it.

Remove remnants from commit id '89144534'

Add checks for existence of GNUTLS and automatically disable
it if not found.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

In a non-systemd environment the post and preun scripts of libvirt-client
fail, since the required files are in libvirt-daemon. Moved them to client.
Doing that I noticed %{_unitdir}/libvirt-guests.service was contained in
both libvirt-client and libvirt-daemon, which I don't think was intended.
Removed the extra copy from daemon.
Signed-off-by: NViktor Mihajlovski <mihajlov@linux.vnet.ibm.com>

Added 'usbFreeDevice' to the useless_free_options list in cfg.mk

gcc 4.1.2 on RHEL 5 warned:
conf/network_conf.c:3136: warning: 'foundIdx' may be used uninitialized in this function

The warning is spurious, but initializing the variable doesn't hurt.

* src/conf/network_conf.c (virNetworkDefUpdateDNSHost): Silence
unused variable warning.

POSIX does not guarantee whether uid_t and gid_t are signed or
unsigned, nor does it guarantee whether they are smaller, same
size, or larger than int (or even the same size as one another).
Therefore, it is possible to have platforms where '(uid_t)-1==-1'
is false or where 'uid = gid = -1' sets uid to the wrong value,
thanks to integer promotion rules.  The only portable way to use
the placeholder value of these two types is to always use a cast.
Thankfully, the issue is mostly theoretical - sanlock only
compiles on Linux for now, and on Linux, these types do not
suffer from strange promotion problems.

* src/locking/lock_driver_sanlock.c
(virLockManagerSanlockSetupLockspace, virLockManagerSanlockInit)
(virLockManagerSanlockCreateLease): Cast -1 to proper type before
comparing with uid_t or gid_t.

Currently, if there's no hard memory limit defined for a domain,
libvirt tries to calculate one, based on domain definition and magic
equation and set it upon the domain startup. The rationale behind was,
if there's a memory leak or exploit in qemu, we should prevent the
host system trashing. However, the equation was too tightening, as it
didn't reflect what the kernel counts into the memory used by a
process. Since many hosts do have a swap, nobody hasn't noticed
anything, because if hard memory limit is reached, process can
continue allocating memory on a swap. However, if there is no swap on
the host, the process gets killed by OOM killer. In our case, the qemu
process it is.

To prevent this, we need to relax the hard RSS limit. Moreover, we
should reflect more precisely the kernel way of accounting the memory
for process. That is, even the kernel caches are counted within the
memory used by a process (within cgroups at least). Hence the magic
equation has to be changed:

  limit = 1.5 * (domain memory + total video memory) + (32MB for cache
          per each disk) + 200MB

Fix copy-paste error doing handshake.

The clientShake was not set to true, thus we'd potentially never leave
the handshake while loop.

Use regfree instead of VIR_FREE.

This is the QEMU backend code for the SCLP console support.
It includes SCLP capability detection, QEMU command line generation
and a test case.
Signed-off-by: NJ.B. Joret <jb@linux.vnet.ibm.com>
Signed-off-by: NViktor Mihajlovski <mihajlov@linux.vnet.ibm.com>

The SCLP console is the native console type for s390 and is preferred
over the virtio console as it doesn't require special drivers and
is more efficient. Recent versions of QEMU come with SCLP support
which is hereby enabled.

The new target types 'sclp' and 'sclplm' can be used to specify a
SCLP console. Adding documentation, domain schema and XML processing
support.
Signed-off-by: NJ.B. Joret <jb@linux.vnet.ibm.com>
Signed-off-by: NViktor Mihajlovski <mihajlov@linux.vnet.ibm.com>

To avoid confusion between the LXC driver <-> controller
monitor RPC protocol and the libvirt-lxc.so <-> libvirtd public
RPC protocol, rename the former to lxc_monitor_protocol.x
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Add some APIs for acquiring namespace file descriptors and
switching namespaces
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Currently the libvirt client can pass FDs to the server, but the
dispatch mechanism provides no way to return FDs back from the
server to the client. Tweak the dispatch code, such that if a
dispatcher returns '1', this indicates that it populated the
virNetMessagePtr with FDs to return
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

A number of bugs handling file descriptors received from the
server caused the FDs to be lost and leaked.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

The QEMU specific APIs all operate on domains, not the host,
so should be in the virsh-domain.c file / group
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Don't dereference 'model' in PowerPCBaseline when there's no outputModel

If running on older Linux without mounted cgroups then its possible that
*root would be NULL.

When virCapabilitiesNew() fails, caps will be NULL resulting in possible
core when deref'd in cpuDataFree() call.

Change calling sequence to only call xenUnifiedDomainSetVcpusFlags() when
'dom' is not NULL. Use the GET_PRIVATE() macro to reference privateData.
Just return -1 if dom is NULL.

The virLXCPrepareHostDevices method was returning success even
when it reported an error, and failed to handle several host
device types

The code for setting up a private /dev/pts for the containers
is also responsible for making the LXC controller have a
private mount namespace. Unfortunately the /dev/pts code is
not run if launching a container without a custom root. This
causes the LXC FUSE mount to leak into the host FS.

Since we daemonized QEMU for capabilities probing there is a long
time if QEMU fails to launch. This is because we're not passing in
any virDomainObjPtr instance and thus the monitor code can not
check to see if the PID is still alive.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

The current code is initializing capabilities before setting
directory permissions. Thus the QEMU binaries being run may
not have the ability to create the UNIX monitor socket on
the first run of libvirtd.

Commit dfa1e1dd removed libxenctrl from LIBXL_LIBS, but the libxl
driver uses a symbol from this library.  Explicitly link with
libxenctrl instead of relying on the build system to support
implicit DSO linking.

https://bugzilla.redhat.com/show_bug.cgi?id=887017 reports that
even though libvirt attempts to set fs.aio-max-nr via sysctl,
the file was installed with the wrong name and gets ignored by
sysctl.  Furthermore, 'man systcl.d' recommends that packages
install into hard-coded /usr/lib/sysctl.d (even when libdir is
/usr/lib64), so that sysadmins can use /etc/sysctl.d for overrides.

* daemon/Makefile.am (install-sysctl, uninstall-sysctl): Use
correct location.
* libvirt.spec.in (network_files): Reflect this.

See also commit 66ff2ddc, where we avoided installing these files
as executables.

* daemon/Makefile.am (libvirtd.service): Drop chmod.
* tools/Makefile.am (libvirt-guests.service): Likewise.
* src/Makefile.am (virtlockd.service, virtlockd.socket):
Likewise.

virtlockd.service could be installed to a configurable root,
but virtlockd.socket was hardcoded to installation into a
distro.

* src/Makefile.am (virtlockd.service, virtlockd.socket): Drop
unused substitutions.
* src/locking/virtlockd.socket.in (ListenStream): Don't hard-code
/var.

We had several different styles of .in conversion in our Makefiles:
ALLCAPS, @ALLCAPS@, @lower@, ::lower::
Canonicalize on one form, to make it easier to copy and paste
between .in files.

Also, we were using some non-portable sed constructs: \@ is an
undefined escape sequence (it happens to be @ itself in GNU sed,
but POSIX allows it to mean something else), as well as risky
behavior (failure to consistently quote things means a space
in $(sysconfdir) could throw things off; also, Autoconf recommends
using | rather than , or ! in the s||| operator, because | has to
be quoted in shell and is therefore less likely to appear in file
names than , or !).

Fix all of these uses to follow the same syntax.

* daemon/libvirtd.8.in: Switch to @var@.
* tools/virt-xml-validate.in: Likewise.
* tools/virt-pki-validate.in: Likewise.
* src/locking/virtlockd.init.in: Likewise.
* daemon/Makefile.am: Prefer | over ! in sed.
(libvirtd.8): Prefer consistent substitution.
(libvirtd.init, libvirtd.service): Avoid non-portable sed.
* tools/Makefile.am (libvirt-guests.sh, libvirt-guests.init)
(libvirt-guests.service): Likewise.
(virt-xml-validate, virt-pki-validate, virt-sanlock-cleanup):
Prefer consistent capitalization.
* src/Makefile.am (virtlockd.init, virtlockd.service)
(virtlockd.socket): Prefer consistent substitution.

This prevents domain starting and disk attaching if the shared disk's
setting conflicts with other active domain(s), E.g. A domain with
"sgio" set as "filtered", however, another active domain is using
it set as "unfiltered".

This ignores the default "filtered" if unpriv_sgio is not supported
by kernel, but for explicit request "filtered", it error out for
domain starting.

Like "rawio", "sgio" is only allowed for block disk of device
type "lun".

It doesn't default disk->sgio to "filtered" when parsing, as
it won't be able to distinguish explicitly requested "filtered"
and a default "filtered" in driver then. We have to error out for
explicit request when the kernel doesn't support the new sysfs
knob "unpriv_sgio", however, for defaulted "filtered", we can
just ignore it if the kernel doesn't support "unpriv_sgio".