Let's fix this before we bake in a painful API.  Since we know
that we have exactly one non-negative fd on success, we might
as well return the fd directly instead of forcing the user to
pass in a pointer.  Furthermore, I found some memory and fd
leaks while reviewing the code - the idea is that on success,
libvirtd will have handed two fds in two different directions:
one to qemu, and one to the RPC client.

* include/libvirt/libvirt.h.in (virDomainOpenGraphicsFD): Drop
unneeded parameter.
* src/driver.h (virDrvDomainOpenGraphicsFD): Likewise.
* src/libvirt.c (virDomainOpenGraphicsFD): Adjust interface to
return fd directly.
* daemon/remote.c (remoteDispatchDomainOpenGraphicsFd): Adjust
semantics.
* src/qemu/qemu_driver.c (qemuDomainOpenGraphicsFD): Likewise,
and plug fd leak.
* src/remote/remote_driver.c (remoteDomainOpenGraphicsFD):
Likewise, and plug memory and fd leak.
Signed-off-by: NEric Blake <eblake@redhat.com>

This commit (finally) adds the virDomainBlockCopy API, with the
intent that it will provide more power to the existing 'virsh
blockcopy' command.

'virsh blockcopy' was first added in Apr 2012 (v0.9.12), which
corresponds to the upstream qemu 1.2 timeframe.  It was done as
a hack on top of the existing virDomainBlockRebase() API call,
for two reasons: 1) it was targetting a feature that landed first
in downstream RHEL qemu, but had not stabilized in upstream qemu
at the time (and indeed, 'drive-mirror' only landed upstream in
qemu 1.3 with slight differences to the first RHEL attempt,
and later gained further parameters like granularity and buf-size
that are also worth exposing), and 2) extending an existing API
allowed it to be backported without worrying about bumping .so
versions.  A virDomainBlockCopy() API was proposed at that time
[1], but we decided not to accept it into libvirt until after
upstream qemu stabilized, and it ended up getting scrapped.
Whether or not RHEL should have attempted adding a new feature
without getting it upstream first is a debate that can be held
another day; but enough time has now elapsed that we are ready to
do the interface cleanly.

[1] https://www.redhat.com/archives/libvir-list/2012-April/msg00768.html

Delaying the creation of a clean API until now has also had a
benefit: we've only recently learned of a few shortcomings in the
original design: 1) it is unable to target a network destination
(such as a gluster volume) because it hard-coded the assumption
that the destination is a local file name.  Because of all the
refactoring we've done to add virStorageSourcePtr, we are in a
better position to declare an API that parses XML describing a
host storage source as the copy destination, which was not
possible had we implemented virDomainBlockCopy as it had been
originally envisioned (although a network target will have to wait
until a later libvirt release compared to the API addition to
actually be implemented).  2) the design of using MiB/sec as the
bandwidth throttle is rather coarse; qemu is actually tuned to
bytes/second, and libvirt is preventing access to that level of
detail.  A later patch will add flags to existing block job API
that can request bytes/second instead of back-compat MiB/s, but as
this is a new API, we can get it right to begin with.

At least I had the foresight to create 'virsh blockcopy' as a
separate command at the UI level (commit 1f06c007) rather than
leaking the underlying API overload of virDomainBlockRebase onto
shell users.

A further note on the bandwidth option: virTypedParameters
intentionally lacks unsigned long (since variable-width
interaction between mixed 32- vs. 64-bit client/server setups is
nasty), but we have to deal with the fact that we are interacting
with existing older code that mistakenly chose unsigned long
bandwidth at a point before we decided to prohibit it in all new
API.  The typed parameter is therefore unsigned long long, but
the implementation (in a later patch) will have to do overflow
detection on 32-bit platforms, as well as capping the value to
match the LLONG_MAX>>20 cap of the existing MiB/s interfaces.

* include/libvirt/libvirt.h.in (virDomainBlockCopy): New API.
(virDomainBlockJobType, virConnectDomainEventBlockJobStatus):
Update related documentation.
* src/libvirt.c (virDomainBlockCopy): Implement it.
* src/libvirt_public.syms (LIBVIRT_1.2.8): Export it.
* src/driver.h (_virDriver): New driver callback.
Signed-off-by: NEric Blake <eblake@redhat.com>

The motivation for this API is that management layers that use libvirt
usually poll for statistics using various split up APIs we currently
provide. To get all the necessary stuff, the app needs to issue a lot of
calls and aggregate the results.

The APIs I'm introducing here:
1) Returns data in a format that we can expand in the future and is
(pseudo) hierarchical. The data is returned as typed parameters where
the fields are constructed as dot-separated strings containing names and
other stuff in a list of typed params.

2) Stats for multiple (all) domains can be queried at once and are
returned in one call. This will decrease the overhead necessary to issue
multiple calls per domain multiplied by the count of domains.

3) Selectable (bit mask) fields in the returned format. This will allow
to retrieve only specific stats according to the app's need.

The stats groups will be enabled using a bit field @stats passed as the
function argument. A few sample stats groups that this API will support:

VIR_DOMAIN_STATS_STATE
VIR_DOMAIN_STATS_CPU
VIR_DOMAIN_STATS_BLOCK
VIR_DOMAIN_STATS_INTERFACE

(Note that this is only an example, the initial implementation supports
 only VIR_DOMAIN_STATS_STATE while others will be added later.)

the returned typed params will use the following scheme

state.state = VIR_DOMAIN_RUNNING
state.reason = VIR_DOMAIN_RUNNING_BOOTED (the actual values according to
                                          the enum)
cpu.count = 8
cpu.0.state = running
cpu.0.time = 1234

Should fix https://bugzilla.redhat.com/show_bug.cgi?id=999926

Define the public API implementation and declare internal
driver prototype.

According to docs/schemas/domaincommon.rng and _virDomainBlockIoTuneInfo
all the iotune values are interpreted as unsigned long long, however
according to qemu_monitor_json.c, qemu silently truncates numbers
larger than LLONG_MAX. There's really not much of a usage for such
large numbers anyway yet. This patch provides the same overflow
check during a domain start as it does during setting
a blkdeviotune element in qemu_driver.c and thus reports an error when
a larger number than LLONG_MAX is detected.

https://bugzilla.redhat.com/show_bug.cgi?id=1131876

Introduced by commit 395171f8.
Signed-off-by: NChen Fan <chen.fan.fnst@cn.fujitsu.com>
Signed-off-by: NJán Tomko <jtomko@redhat.com>

QEMU 2.1 added support for the kvm=off option to the -cpu command,
allowing the KVM hypervisor signature to be hidden from the guest.
This enables disabling of some paravirualization features in the
guest as well as allowing certain drivers which test for the
hypervisor to load.  Domain XML syntax is as follows:

<domain type='kvm>
  ...
  <features>
    ...
    <kvm>
      <hidden state='on'/>
    </kvm>
  </features>
  ...
Signed-off-by: NAlex Williamson <alex.williamson@redhat.com>

Commit b55cc5f4 did a shallow copy of libxl_{sdl,vnc}_info from the
domain config to the build info, which resulted in double-freeing
strings contained in the structures during cleanup, which later
resulted in a libvirtd crash.  Fix by performing a deep copy of the
structure, VIR_STRDUP'ing embedded strings instead of simply copying
their pointers.

Fixes the following issue reported on the libvirt dev list

https://www.redhat.com/archives/libvir-list/2014-August/msg01112.html

I noticed a line 'int nparams = 0;;' in remote_dispatch.h, and
tracked down where it was generated.  While at it, I found a
couple of other double semicolons.  Additionally, I noticed that
commit df0b57a9 left a stale reference to the file name
remote_dispatch_bodies.h.

* src/conf/numatune_conf.c (virDomainNumatuneNodeParseXML): Drop
empty statement.
* tests/virdbustest.c (testMessageStruct, testMessageSimple):
Likewise.
* src/rpc/gendispatch.pl (remote_dispatch_bodies.h): Likewise, and
update stale comments.
Signed-off-by: NEric Blake <eblake@redhat.com>

When trying to set an invalid value into iotune element, standard
behavior was to not report any error, rather to reset all affected
subelements of the iotune element back to 0 which results in ignoring
those particular subelements by XML generator. Patch further
examines the return code of the virXPathULongLong function
and in case of an invalid non-integer value raises an error.
Fixed to preserve consistency with invalid value checking
of other elements.

Resolves https://bugzilla.redhat.com/show_bug.cgi?id=1131811

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1021703Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

The commit "f5b4c141" introduced new "force" parameter
for "virFDStreamOpenFileInternal" but forget to update
one call of that function.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

Add an implementation of uploadVol and downloadVol using
virStorageBackendVolUploadLocal and virStorageBackendVolDownloadLocal
respectively.

virStorageBackendVolDownloadLocal and virStorageBackendVolUploadLocal
use virFDStreamOpenFile function to work with the volume fd.

virFDStreamOpenFile calls virFDStreamOpenFileInternal that implements
handling of the non-blocking I/O. If a file is not a character device and
not a fifo, it uses libvirt_iohelper.

On FreeBSD, it doesn't work as expected because disk devices (including
ZFS volumes) are exposed as character devices, and ZFS volumes do not
support open(2) with O_NONBLOCK.

To overcome this, introduce a forceIOHelper flag to
virFDStreamOpenFileInternal that forces using libvirt_iohelper. And
introduce virFDStreamOpenBlockDevice that calls
virFDStreamOpenFileInternal with the forceIOHelper set to true.

virFDStreamOpenInternal terminates if virSetNonBlock fails. As
virSetNonBlock uses gnulib's set_nonblocking_flag that sets errno,
call virReportSystemError() to let user know the reason of fail.

* src/util/virbuffer.h: s/occured/occurred/
Signed-off-by: NEric Blake <eblake@redhat.com>

Commit b606bbb4 reminded me that any time we drop locks to run
back-to-back guest interaction commands, we have to check that
the guest didn't disappear in between the two commands.  A quick
audit found a couple of spots that were missing this check.

* src/qemu/qemu_driver.c (qemuDomainShutdownFlags)
(qemuDomainSetVcpusFlags): Check that domain is still up.
Signed-off-by: NEric Blake <eblake@redhat.com>

Since '337a1362' - Coverity complains that 'net' is VIR_ALLOC()'d, but
on various 'cleanup' exit paths from the code there is no corresponding
cleanup.

Since '1b807f92' - Coverity complains that in the error paths of
both virFork() and virProcessWait() that the 'passfd' will not be closed.
Added the VIR_FORCE_CLOSE(passfd) and initialized it to -1.

Also noted that variable 'buf' was never really used - so I removed it

When trying to set numatune mode directly using virsh numatune command,
correct error is raised, however numatune structure was not deallocated,
thus resulting in creating an empty numatune element in the guest XML,
if none was present before. Running the same command aftewards results
in a successful change with broken XML structure. Patch fixes the
deallocation problem as well as checking for invalid attribute
combination VIR_DOMAIN_NUMATUNE_PLACEMENT_AUTO + a nonempty nodeset.

Resolves https://bugzilla.redhat.com/show_bug.cgi?id=1129998

The 'min_guarantee' is used by VMware ESX and OpenVZ drivers,
with qemu however, libvirt should report error when starting a domain,
because this element is not used.
Resolves https://bugzilla.redhat.com/show_bug.cgi?id=1122455

On some places in the libvirt code we have:

  f(a,z)

instead of

  f(a, z)

This trivial patch fixes couple of such occurrences.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Wrong error code in config errors reporting was used. Fixed it.

Libvirt measures vram in Kbytes, not in bytes, so calculation
of Mbytes was incorrect. PCS server can take vram argument
with units, so I added K postfix to make params a little bit clearer.

That function caused errors in libvirtd logs when OpenStack Nova
starts VM instance.
Signed-off-by: NJán Tomko <jtomko@redhat.com>

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1064770Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

This eliminates the need for active waiting.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=927369Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

That sets a new flag, but that flag does mean the child will get
LISTEN_FDS and LISTEN_PID environment variables properly set and
passed FDs reordered so that it corresponds with LISTEN_FDS (they must
start right after STDERR_FILENO).
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

It's just a wrapper around NewFD and NewUNIX that selects the right
option and increments the number of used FDs.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Since not only systemd can do this (we'll be doing it as well few
patches later), change 'systemd' to 'caller' and fix LISTEN_FDS to
LISTEN_PID where applicable.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

When formatting the forward mode addresses or interfaces the switch was
done based on the type of the network rather than of the type of the
individual <interface>/<address> element. In case a user would specify
an incorrect network type ("passhtrough") with <address> elements,
libvirtd would crash as it would attempt to format an <interface>.

Use the type of the individual element to format the XML.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1132347

https://bugzilla.redhat.com/show_bug.cgi?id=1078126

Using 'virsh attach-device --config' (or --persistent) to attach a
file backed lun device will succeed; however, subsequent domain restarts
will result in failure because the configuration of a file backed lun
is not supported.

Although allowing 'illegal configurations' is something that can be
allowed, it may not be practical in this case. Generally, when attaching
a device to a domain means the domain must be running. A way around
this is using the --config (or --persistent) option. When an attach
is done to a running domain, a temporary configuration is modified
first followed by the live update. The live update will make a number
of disk validity checks when building the qemu command to attach the
disk. If any fail, then change is rejected.

Rather than allow a potentially illegal combination, adjust the code
in the configuration path to make the same checks as the running path
will make with respect to disk validity checks. This way we avoid
having the potential for some subsequent start/reboot to fail because
an illegal combination was allowed.

NB: The live path still checks the configuration since it is possible
to just do --live guest modification...

Since vbox driver rewrite the virDriver structure init moved from
vbox_tmpl.c into vbox_common.c. However, our hvsupport.pl script
doesn't count with that. It still parses vbox_tmp.c and looks for
virDriver structure which is not found there anymore. As a result,
at hvsupport page is seems like vbox driver doesn't support
anything.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

In case the host has 2 or more NUMA nodes, we fetch CPU map for each
node. However, we need to free the CPU map in between loops:

==29513== 96 (72 direct, 24 indirect) bytes in 3 blocks are definitely lost in loss record 951 of 1,264
==29513==    at 0x4C2A700: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==29513==    by 0x52AD24B: virAlloc (viralloc.c:144)
==29513==    by 0x52AF0E6: virBitmapNew (virbitmap.c:78)
==29513==    by 0x52FB720: virNumaGetNodeCPUs (virnuma.c:294)
==29513==    by 0x53C700B: nodeCapsInitNUMA (nodeinfo.c:1886)
==29513==    by 0x11759708: vboxCapsInit (vbox_common.c:398)
==29513==    by 0x11759CC4: vboxConnectOpen (vbox_common.c:514)
==29513==    by 0x53C965F: do_open (libvirt.c:1147)
==29513==    by 0x53C9EBC: virConnectOpen (libvirt.c:1317)
==29513==    by 0x142905: remoteDispatchConnectOpen (remote.c:1215)
==29513==    by 0x126ADF: remoteDispatchConnectOpenHelper (remote_dispatch.h:2346)
==29513==    by 0x5453D21: virNetServerProgramDispatchCall (virnetserverprogram.c:437)
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>