Another layer of overly-multiplexed code that deserves to be
split into obviously separate paths for query vs. modify.
This continues the cleanup started in commit cefe0ba3.

In the process, make some tweaks to simplify the logic when
parsing the JSON reply.  There should be no user-visible
semantic changes.

* src/qemu/qemu_monitor.h (qemuMonitorBlockJob): Drop parameter.
(qemuMonitorBlockJobInfo): New prototype.
(BLOCK_JOB_INFO): Drop enum.
* src/qemu/qemu_monitor_json.h (qemuMonitorJSONBlockJob)
(qemuMonitorJSONBlockJobInfo): Likewise.
* src/qemu/qemu_monitor.c (qemuMonitorBlockJob): Split...
(qemuMonitorBlockJobInfo): ...into second function.
* src/qemu/qemu_monitor_json.c (qemuMonitorJSONBlockJob): Move
block info portions...
(qemuMonitorJSONGetBlockJobInfo): ...here, and rename...
(qemuMonitorJSONBlockJobInfo): ...and export.
(qemuMonitorJSONGetBlockJobInfoOne): Alter return semantics.
* src/qemu/qemu_driver.c (qemuDomainBlockPivot)
(qemuDomainBlockJobImpl, qemuDomainGetBlockJobInfo): Adjust
callers.
* src/qemu/qemu_migration.c (qemuMigrationDriveMirror)
(qemuMigrationCancelDriveMirror): Likewise.
Signed-off-by: NEric Blake <eblake@redhat.com>

The qemu implementation for virDomainGetBlockJobInfo() has a
minor bug: it grabs the qemu job with intent to QEMU_JOB_MODIFY,
which means it cannot be run in parallel with any other
domain-modifying command.  Among others, virDomainBlockJobAbort()
is such a modifying command, and it defaults to being
synchronous, and can wait as long as several seconds to ensure
that the job has actually finished.  Due to the job rules, this
means a user cannot obtain status about the job during that
timeframe, even though we know that some client management code
exists which is using a polling loop on status to see when a job
finishes.

This bug has been present ever since blockpull support was first
introduced (commit b976165c, v0.9.4 in Jul 2011), all because we
stupidly tried to cram too much multiplexing through a single
helper routine, but was made worse in 97c59b9c (v1.2.7) when
BlockJobAbort was fixed to wait longer.  It's time to disentangle
some of the mess in qemuDomainBlockJobImpl, and in the process
relax block job query to use QEMU_JOB_QUERY, since it can safely
be used in parallel with any long running modify command.

Technically, there is one case where getting block job info can
modify domain XML - we do snooping to see if a 2-phase job has
transitioned into the second phase, for an optimization in the
case of old qemu that lacked an event for the transition.  I
claim this optimization is safe (the jobs are all about modifying
qemu state, not necessarily xml state); but if it proves to be
a problem, we could use the difference between the capabilities
QEMU_CAPS_BLOCKJOB_{ASYNC,SYNC} to determine whether we even
need snooping, and only request a modifying job in the case of
older qemu.

* src/qemu/qemu_driver.c (qemuDomainBlockJobImpl): Move info
handling...
(qemuDomainGetBlockJobInfo): ...here, and relax job type.
(qemuDomainBlockJobAbort, qemuDomainBlockJobSetSpeed)
(qemuDomainBlockRebase, qemuDomainBlockPull): Adjust callers.
Signed-off-by: NEric Blake <eblake@redhat.com>

The existing virDomainBlockRebase code rejected the combination of
_RELATIVE and _COPY flags, but only by accident.  It makes sense
to add support for the combination someday, at least for the case
of _SHALLOW and not _REUSE_EXT; but to implement it, libvirt would
have to pre-create the file with a relative backing name, and I'm
not ready to code that in yet.

Meanwhile, the code to forward on to the block copy code is getting
longer, and reorganizing the function to have the block pull done
early makes it easier to add even more block copy prep code.

This patch should have no semantic difference other than the quality
of the error message on the unsupported flag combination.  Pre-patch:

error: unsupported flags (0x10) in function qemuDomainBlockCopy

Post-patch:

error: argument unsupported: Relative backing during copy not supported yet

* src/qemu/qemu_driver.c (qemuDomainBlockRebase): Reorder code,
and improve error message of relative copy.
Signed-off-by: NEric Blake <eblake@redhat.com>

I'm about to add a syntax check that enforces our documented
HACKING style of always using matching {} on if-else statements.

This commit focuses on the qemu driver.

* src/qemu/qemu_command.c (qemuParseISCSIString)
(qemuParseCommandLineDisk, qemuParseCommandLine)
(qemuBuildSmpArgStr, qemuBuildCommandLine)
(qemuParseCommandLineDisk, qemuParseCommandLineSmp): Correct use
of {}.
* src/qemu/qemu_capabilities.c (virQEMUCapsProbeCPUModels):
Likewise.
* src/qemu/qemu_driver.c (qemuDomainCoreDumpWithFormat)
(qemuDomainRestoreFlags, qemuDomainGetInfo)
(qemuDomainMergeBlkioDevice): Likewise.
* src/qemu/qemu_hotplug.c (qemuDomainAttachNetDevice): Likewise.
* src/qemu/qemu_monitor_text.c (qemuMonitorTextCreateSnapshot)
(qemuMonitorTextLoadSnapshot, qemuMonitorTextDeleteSnapshot):
Likewise.
* src/qemu/qemu_process.c (qemuProcessStop): Likewise.
Signed-off-by: NEric Blake <eblake@redhat.com>

If virSecurityManagerClearSocketLabel() fails, 'agent' won't
be freed before jumping to cleanup.
Signed-off-by: NWang Rui <moon.wangrui@huawei.com>

Add umask to _virCommand, allow user to set umask to command.
Set umask(002) to qemu process to overwrite the default umask
of 022 set by many distros, so that unix sockets created for
virtio-serial has expected permissions.

Fix problem reported here:
https://sourceware.org/bugzilla/show_bug.cgi?id=13078#c11
https://bugzilla.novell.com/show_bug.cgi?id=888166

To use virtio-serial device, unix socket created for chardev with
default umask(022) has insufficient permissions.
e.g.:
-device virtio-serial \
-chardev socket,path=/tmp/foo,server,nowait,id=foo \
-device virtserialport,chardev=foo,name=org.fedoraproject.port.0

srwxr-xr-x 1 qemu qemu 0 21. Jul 14:19 /tmp/somefile.sock

Other users in the same group (like real user, test engines, etc)
cannot write to this socket.
Signed-off-by: NChunyan Liu <cyliu@suse.com>
Signed-off-by: NEric Blake <eblake@redhat.com>

Currently, there is one flag passed in during macvtap creation
(withTap) -- Let's convert this field to an unsigned int flag
field for future expansion.
Signed-off-by: NMatthew Rosato <mjrosato@linux.vnet.ibm.com>
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

In qemuDomainSnapshotCreateDiskActive() if we jumped to cleanup from a
failed actions = virJSONValueNewArray(), then 'cfg' would be NULL.

So just return -1, which in turn removes the need for cleanup:

For virtio-blk-pci disks with the disk iothread attribute that are
running the correct emulator, add the "iothread=iothread#" to the
-device command line in order to enable iothreads for the disk as
long as the command is available, the disk iothread value provided is
valid, and is supported for the disk device being added

Add a new capability to ensure the iothreads feature exists for the qemu
emulator being run - requires the "query-iothreads" QMP command. Using the
domain XML add correspoding command argument in order to generate the
threads. The iothreads will use a name space "iothread#" where, the
future patch to add support for using an iothread to a disk definition to
merely define which of the available threads to use.

Add tests to ensure the xml/argv processing is correct.  Note that no
change was made to qemuargv2xmltest.c as processing the -object element
would require knowing more than just iothreads.

Implement the API function for virDomainListGetStats and
virConnectGetAllDomainStats in a modular way and implement the
VIR_DOMAIN_STATS_STATE group of statistics.

Although it may look like the function looks universal I'd rather not
expose it to other drivers as the coming stats groups are likely to do
qemu specific stuff to obtain the stats.

One useless warning, but the other one rather pertinent. On entry
the 'trans' variable is initialized to VIR_DOMAIN_DISK_TRANS_DEFAULT.
When the "trans" was found in the parsing loop it def->geometry.trans
was assigned to the return from virDomainDiskGeometryTransTypeFromString
and then 'trans' was used to do the comparison to see if it was valid.

So remove 'trans' and use def->geometry.trans properly

A bunch of false positives brought on by our own doings

The PROBE macro can expand to more than one line/statement - put curly
braces around the if statement to be safe

Coverity found that the 'buf' wasn't VIR_FREE'd at exit.

Coverity determined that on error path that 'mach' wouldn't be free'd
Since virCapabilitiesFreeGuestMachine() isn't globally available, we'll
insert first and then if the VIR_STRDUP's fail they it will eventually
cause the 'mach' to be freed in the error path

Coverity found that on error paths, the 'arg' value wasn't be cleaned
up. Followed the example in qemuAgentSetVCPUs() where upon successful call
to qemuAgentCommand() the 'cpus' is set to NULL; otherwise, when cleanup
occurs the free the memory for 'arg'

In qemuParseISCSIString() if an error was returned, then the call
to qemuParseDriveURIString() where the uri is free'd wouldn't be run

In function virQEMUCapsParseMachineTypesStr, VIR_STRNDUP allocates
memory for 'name' in {do,while} loop. If 'name' isn't freed before
'continue', its memory will be allocated again in the next loop.
In this case the memory allocated for 'name' in privious loop is
useless and not freed. Free it before continue this loop to fix that.
Signed-off-by: NWang Rui <moon.wangrui@huawei.com>

Adjust the initialization of qemuCaps() to check for a NULL before
attempting to dereference like other callers/users do.

The call to virDomainSnapshotRedefinePrep() had a spurrious ! in front of
it which caused Coverity to complan that the expression is always false.

In qemuNetworkIfaceConnect() a call to virNetDevBandwidthSet() is
made where the function prototype requires the first parameter
(net->ifname) to be non NULL.  Coverity complains that the subsequent
non NULL check for net->ifname prior to the next call gets flagged as
an unnecessary check.  Resolve by removing the extra check

In qemuDomainRevertToSnapshot(), it will check snap->def->state.
But when the state is PMSUSPENDED/NOSTATE/BLOCKED, it forgets to
call qemuDomainObjEndJob.

https://bugzilla.redhat.com/show_bug.cgi?id=1134154
Bug introduced in commit 1e833899.
Signed-off-by: NJincheng Miao <jmiao@redhat.com>
Signed-off-by: NEric Blake <eblake@redhat.com>

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>

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

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

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>

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>

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>

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>

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...

And add a syntax-check for '){$'.  It's not perfect, but better than
nothing.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1103245

An advice appeared there on the qemu-devel list [1]. When a domain is
suspended and then resumed guest kernel is not aware of this. So we've
introduced virDomainSetTime API that resets the time within guest
using qemu-ga. On the other hand, qemu itself is trying to make RTC
beat faster to catch the difference. But if we don't tell qemu that
guest's time was reset via the other method, both mechanisms are
applied resulting in again wrong guest time. In order to avoid summing
both corrections we need to tell qemu that it should not use the RTC
injection if the guest time is set via guest agent.

1: http://www.mail-archive.com/qemu-devel@nongnu.org/msg236435.htmlSigned-off-by: NMichal Privoznik <mprivozn@redhat.com>

Partially resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1131306Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

The build failure is caused by a false positive of some static analysys
steps done by gcc (that don't happen on -O0).

When a user would try changing the persistent IO tuning settings for a
disk that was hotplugged to a vm in a transient way, the
qemuDomainSetBlockIoTune API would use the same index for both the
live and config disk array. The disk was missing from the config array
though causing a crash of libvirtd.

To fix the issue, determine the indexes separately.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1131819

https://bugzilla.redhat.com/show_bug.cgi?id=1095636

When starting up the domain the domain's NICs are allocated. As of
1f24f682 (v1.0.6) we are able to use multiqueue feature on virtio
NICs. It breaks network processing into multiple queues which can be
processed in parallel by different host CPUs. The queues are, however,
created by opening /dev/net/tun several times. Unfortunately, only the
first FD in the row is labelled so when turning the multiqueue feature
on in the guest, qemu will get AVC denial. Make sure we label all the
FDs needed.

Moreover, the default label of /dev/net/tun doesn't allow
attaching a queue:

    type=AVC msg=audit(1399622478.790:893): avc:  denied  { attach_queue }
    for  pid=7585 comm="qemu-kvm"
    scontext=system_u:system_r:svirt_t:s0:c638,c877
    tcontext=system_u:system_r:virtd_t:s0-s0:c0.c1023
    tclass=tun_socket

And as suggested by SELinux maintainers, the tun FD should be labeled
as svirt_t. Therefore, we don't need to adjust any range (as done
previously by Guannan in ae368ebf) rather set the seclabel of the
domain directly.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Instead of tediously copying of the disk source to remove it later
ensure that the media change function removes the old device after it
succeeds.