This patch is rather cosmetic as it only moves device alias
assignation from command line construction just before that.
However, it is needed in connotation of previous and next patch.

Detected by Coverity.  Both text and JSON monitors set only the
bus and unit fields, which means driveAddr.controller spends
life as garbage on the stack, and is then memcpy()'d into the
in-memory representation which the user can see via dumpxml.

* src/qemu/qemu_hotplug.c (qemuDomainAttachSCSIDisk): Only copy
defined fields.

The improvements to virBuffer, along with a paradigm shift to pass
the original buffer through rather than creating a second buffer,
allow us to shave off quite a few lines of code.

* src/util/sysinfo.h (virSysinfoFormat): Alter signature.
* src/util/sysinfo.c (virSysinfoFormat, virSysinfoBIOSFormat)
(virSysinfoSystemFormat, virSysinfoProcessorFormat)
(virSysinfoMemoryFormat): Change indentation parameter.
* src/conf/domain_conf.c (virDomainSysinfoDefFormat): Adjust
caller.
* src/qemu/qemu_driver.c (qemuGetSysinfo): Likewise.

<domainsnapshot> is the first public instance of <domain> being
used as a sub-element, although we have two other private uses
(runtime state, and migration cookie).  Although indentation has
no effect on XML parsing, using it makes the output more consistent.

This uses virBuffer auto-indentation to obtain the effect, for all
but the portions of <domain> that are not generated a line at a
time into the same virBuffer.  Further patches will clean up the
remaining problems.

* src/conf/domain_conf.h (virDomainDefFormatInternal): New prototype.
* src/conf/domain_conf.c (virDomainDefFormatInternal): Export.
(virDomainObjFormat, virDomainSnapshotDefFormat): Update callers.
* src/libvirt_private.syms (domain_conf.h): Add new export.
* src/qemu/qemu_migration.c (qemuMigrationCookieXMLFormat): Use
new function.
(qemuMigrationCookieXMLFormatStr): Update caller.

There is a little difference between the output of domxml-to-native and the actual commandline.
No matter qemu is in control or readline mode, domxml-to-native always converts it to readline mode.

That is because the parameter "monitor_json" for qemuBuildCommandLine() is always set to false
in qemuDomainXMLToNative().
Signed-off-by: Ntangchen <tangchen@cn.fujitsu.com>

The XML parser for the qemu specific extensions expects the qemu name-space
to be bound to the 'qemu' prefix. This is too strict, since the name of the
name-space-prefix is only meant as an internal lookup key. Only the associated
URI is relevant.
<domain>...
  <qemu:commandline xmlns:qemu="http://libvirt.org/schemas/domain/qemu/1.0">
  ...</qemu:commandline>
</domain>

<domain xmlns:ns0="http://libvirt.org/schemas/domain/qemu/1.0">...
  <ns0:commandline>
  ...</ns0:commandline>
</domain>

<domain xmlns:qemu="http://libvirt.org/schemas/domain/qemu/1.0">
  <qemu:commandline xmlns:qemu="urn:foo">
  ...</qemu:commandline>
</domain>

Remove the test for checking the name-space binding on the top-level <domain>
element. Registering the name-space with XPath is enough.
Signed-off-by: NPhilipp Hahn <hahn@univention.de>

Noticed when testing new libvirt against old qemu that lacked the
snapshot_blkdev HMP command.  Libvirt was mistakenly treating the
command as successful, and re-writing the domain XML to use the
just-created 0-byte file, rendering the domain broken on restart.

* src/qemu/qemu_monitor_text.c (qemuMonitorTextDiskSnapshot):
Notice another possible error message.
* src/qemu/qemu_driver.c
(qemuDomainSnapshotCreateSingleDiskActive): Don't keep 0-byte file
on failure.

Waiting for qemu-kvm cleaning up the PCI bar(s) mapping with long time
while the device is not managed is just waste of time.

Tested with multiple AHCI controllers and multiple disks attached
to a controller. E.g.,

    <disk type='file' device='disk'>
      <driver name='qemu' type='raw'/>
      <source file='/var/lib/libvirt/images/test/disk0.raw'/>
      <target dev='sda' bus='sata'/>
      <address type='drive' controller='0' bus='0' unit='0'/>
    </disk>
    <disk type='file' device='disk'>
      <driver name='qemu' type='raw'/>
      <source file='/var/lib/libvirt/images/test/disk1.raw'/>
      <target dev='sdb' bus='sata'/>
      <address type='drive' controller='0' bus='0' unit='1'/>
    </disk>
    <disk type='file' device='disk'>
      <driver name='qemu' type='raw'/>
      <source file='/var/lib/libvirt/images/test/disk2.raw'/>
      <target dev='sdc' bus='sata'/>
      <address type='drive' controller='1' bus='0' unit='0'/>
    </disk>
    <controller type='sata' index='0'>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/>
    </controller>
    <controller type='sata' index='1'>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/>
    </controller>

The patch that fixes SIGTERM handling with -no-shutdown was taken into
0.15.1 stable release of qemu.

BZ# https://bugzilla.redhat.com/show_bug.cgi?id=736214

The problem is caused by the original info of domain's PCI dev is
maintained by qemu_driver->activePciHostdevs list, (E.g. dev->reprobe,
which stands for whether need to reprobe driver for the dev when do
reattachment). The fields (dev->reprobe, dev->unbind_from_stub, and
dev->remove_slot) are initialized properly when preparing the PCI
device for managed attachment. However, when do reattachment, it
construct a complete new "pciDevice" without honoring the original
dev info, and thus the dev won't get the original driver or can get
other problem.

This patch is to fix the problem by get the devs from list
driver->activePciHostdevs.

Tested with following 3 scenarios:
  * the PCI was bound to some driver not pci-stub before attaching

    result: the device will be bound to the original driver

  * the PCI was bound to pci-stub before attaching

    result: no driver reprobing, and still bound to pci-stub

  * The PCI was not bound to any driver

    result: no driver reprobing, and still not bound to any driver.

Explicitly disallow conflicts between domain name from dxml and dname.

When failing on starting a domain, it tries to reattach all the PCI
devices defined in the domain conf, regardless of whether the devices
are still used by other domain. This will cause the devices to be deleted
from the list qemu_driver->activePciHostdevs, thus the devices will be
thought as usable even if it's not true. And following commands
nodedev-{reattach,reset} will be successful.

How to reproduce:
  1) Define two domains with same PCI device defined in the confs.
  2) # virsh start domain1
  3) # virsh start domain2
  4) # virsh nodedev-reattach $pci_device

You will see the device will be reattached to host successfully.
As pciDeviceReattach just check if the device is still used by
other domain via checking if the device is in list driver->activePciHostdevs,
however, the device is deleted from the list by step 2).

This patch is to prohibit the bug by:
  1) Prohibit a domain starting or device attachment right at
     preparation period (qemuPrepareHostdevPCIDevices) if the
     device is in list driver->activePciHostdevs, which means
     it's used by other domain.

  2) Introduces a new field for struct _pciDevice, (const char *used_by),
     it will be set as the domain name at preparation period,
     (qemuPrepareHostdevPCIDevices). Thus we can prohibit deleting
     the device from driver->activePciHostdevs if it's still used by
     other domain when stopping the domain process.

* src/pci.h (define two internal functions, pciDeviceSetUsedBy and
    pciDevceGetUsedBy)
* src/pci.c (new field "const char *used_by" for struct _pciDevice,
    implementations for the two new functions)
* src/libvirt_private.syms (Add the two new internal functions)
* src/qemu_hostdev.h (Modify the definition of functions
    qemuPrepareHostdevPCIDevices, and qemuDomainReAttachHostdevDevices)
* src/qemu_hostdev.c (Prohibit preparation and don't delete the
    device from activePciHostdevs list if it's still used by other domain)
* src/qemu_hotplug.c (Update function usage, as the definitions are
    changed)
Signed-off-by: NEric Blake <eblake@redhat.com>

Detected by Coverity.  Leak introduced in commit 72de0d28.

* src/qemu/qemu_migration.c (qemuMigrationCookieGraphicsXMLParse):
Clean up on success.

Otherwise we can end up with a dangling job that can only be cleared by
restarting libvirtd.

Log debug messages anytime we call *BeginJob* or *EndJob* so that it's
easier to spot incorrect usage of domain job APIs.

Detected by Coverity.  p (the pointer to the string) is always true;
when in reality, we wanted to know whether the integer value of the
just-parsed string is '0' or '1'.  Logic bug since commit b1b5b51a.

* src/qemu/qemu_monitor_text.c (qemuMonitorTextGetBlockInfo): Set
results to proper value.

Detected by Coverity.  If, for some reason, our text monitor input
does not match our assumptions, we end up incrementing p while it
is NULL, then dereferencing the pointer 0x1, which will fault.

* src/qemu/qemu_monitor_text.c
(qemuMonitorTextGetBlockStatsParamsNumber): Rewrite to avoid
deref of strchr failure.  Fix indentation.

Detected by Coverity.  Introduced in commit b1b5b51a.

* src/qemu/qemu_monitor_json.c (qemuMonitorJSONGetBlockInfo):
Avoid null dereference.

As this is needed. Although some functions check for domain
being active before obtaining job, we need to check it after,
because obtaining job unlocks domain object, during which
a state of domain can be changed.

This patch extends qemudDomainCoreDump so it supports new VIR_DUMP_RESET
flag. If this flag is set, domain is reset on successful dump. However,
this is needed to be done after we start CPUs.

With the recent refactoring of qemu snapshot relationships, it
is now trivial to filter on leaves.

* src/conf/domain_conf.c (virDomainSnapshotObjListCount)
(virDomainSnapshotObjListCopyNames): Handle new flag.
* src/qemu/qemu_driver.c (qemuDomainSnapshotListNames)
(qemuDomainSnapshotNum, qemuDomainSnapshotListChildrenNames)
(qemuDomainSnapshotNumChildren): Pass new flag through.

VirtFS allows the user to choose between path/handle based fs driver.
As of now, libvirt hardcoded path based driver only. This patch provides
a solution to allow user to choose between path/handle based fs driver.

Sample:

    <filesystem type='mount'>
      <driver type='handle'/>
      <source dir='/folder/to/share1'/>
      <target dir='mount_tag1'/>
    </filesystem>

    <filesystem type='mount'>
      <driver type='path'/>
      <source dir='/folder/to/share2'/>
      <target dir='mount_tag2'/>
    </filesystem>
Signed-off-by: NHarsh Prateek Bora <harsh@linux.vnet.ibm.com>
Signed-off-by: NEric Blake <eblake@redhat.com>

The previous optimizations lead to some follow-on cleanups.

* src/conf/domain_conf.c (virDomainSnapshotForEachChild)
(virDomainSnapshotForEachDescendant): Drop dead parameter.
(virDomainSnapshotActOnDescendant)
(virDomainSnapshotObjListNumFrom)
(virDomainSnapshotObjListGetNamesFrom): Update callers.
* src/qemu/qemu_driver.c (qemuDomainSnapshotNumChildren)
(qemuDomainSnapshotListChildrenNames, qemuDomainSnapshotDelete):
Likewise.
* src/conf/domain_conf.h: Update prototypes.

Maintain the parent/child relationships of all qemu snapshots.

* src/qemu/qemu_driver.c (qemuDomainSnapshotLoad): Populate
relationships after loading.
(qemuDomainSnapshotCreateXML): Set relations on creation; tweak
redefinition to reuse existing object.
(qemuDomainSnapshotReparentChildren, qemuDomainSnapshotDelete):
Clear relations on delete.

To date, JSON disk snapshots worked by accident, as they were always
using hmp fallback due to a typo in commit e702b5ba not picking up
on the (intentional) difference in command names between the two
monitor protocols.

* src/qemu/qemu_monitor_json.c (qemuMonitorJSONDiskSnapshot):
Spell QMP command correctly.
Reported by Luiz Capitulino.

Not too hard to wire up.  The trickiest part is realizing that
listing children of a snapshot cannot use SNAPSHOT_LIST_ROOTS,
and that we overloaded that bit to also mean SNAPSHOT_LIST_DESCENDANTS;
we use that bit to decide which iteration to use, but don't want
the existing counting/listing functions to see that bit.

* src/conf/domain_conf.h (virDomainSnapshotObjListNumFrom)
(virDomainSnapshotObjListGetNamesFrom): New prototypes.
* src/conf/domain_conf.c (virDomainSnapshotObjListNumFrom)
(virDomainSnapshotObjListGetNamesFrom): New functions.
* src/libvirt_private.syms (domain_conf.h): Export them.
* src/qemu/qemu_driver.c (qemuDomainSnapshotNumChildren)
(qemuDomainSnapshotListChildrenNames): New functions.

If parsing qemu command line fails (e.g. because of non-existing
process number supplied), we jump to cleanup label where we free
pidfile. Therefore it needs to be initialized. Otherwise we free
random pointer.

Coverity complained that 4 out of 5 callers to virJSONValueObjectGetBoolean
checked for errors.  But we documented that we don't care in this case.

* src/qemu/qemu_monitor_json.c (qemuMonitorJSONGetBlockInfo): Use
ignore_value.

Previously libvirt's disk device XML only had a single attribute,
error_policy, to control both read and write error policy, but qemu
has separate options for controlling read and write. In one case
(enospc) a policy is allowed for write errors but not read errors.

This patch adds a separate attribute that sets only the read error
policy. If just error_policy is set, it will apply to both read and
write error policy (previous behavior), but if the new rerror_policy
attribute is set, it will override error_policy for read errors only.
Possible values for rerror_policy are "stop", "report", and "ignore"
("report" is the qemu-controlled default for rerror_policy when
error_policy isn't specified).

For consistency, the value "report" has been added to the possible
values for error_policy as well.

commit 12062abb set rerror=ignore when error_policy="enospace" was
selected (since the rerror option in qemu doesn't accept "enospc", as
the werror option does).

After that patch was already pushed, Paolo Bonzini noticed it and
commented that leaving rerror at the default ("report") would be a
better choice. This patch corrects the problem - if error_policy =
"enospace" is given, rerror is left off the qemu commandline,
effectively setting it to "report". For other values, rerror is still
set to match werror.

Additionally, the parsing of error_policy was changed to no longer
erroneously allow "default" as a choice - as with most other
attributes, if you want the default setting, just don't specify an
error_policy.

Finally, two ommissions in the first patch were corrected - a
long-dormant qemuxml2argv test for enospace was enabled, and fixed to
pass, and the argv2xml parser in qemu_command.c was updated to
recognize the different spelling on the qemu commandline.

Now that RHEL 6.2 Beta is out, it would be nice to test multifunction
devices on that platform.  This changes things so that the multifunction
cap bit can be set in two different ways: by version comparison (needed
for qemu 0.13 which lacked a -device query), and by -device query
(provided by qemu.git and backported to the RHEL beta build of
qemu-kvm which still claims to be a modified 0.12, and therefore needed
for RHEL).

* src/qemu/qemu_capabilities.c (qemuCapsParseDeviceStr): Allow
second method of setting multifunction cap bit.
* tests/qemuhelptest.c (mymain): Test it.
* tests/qemuhelpdata/qemu-kvm-0.12.1.2-rhel62-beta: New file.
* tests/qemuhelpdata/qemu-kvm-0.12.1.2-rhel62-beta-device: Likewise.

Implements the documentation for snapshot revert vs. force.

Part of the patch tightens existing behavior (previously, reverting
to an old snapshot without <domain> was blindly attempted, now it
requires force), while part of it relaxes behavior (previously, it
was not possible to revert an active domain to an ABI-incompatible
active snapshot, now force allows this transition).

* src/qemu/qemu_driver.c (qemuDomainRevertToSnapshot): Check for
risky situations, and allow force to get past them.

Once we know which set of disks belong to a snapshot, reverting or
deleting that snapshot should visit just those disks, rather than
also visiting disks that were hot-plugged in the meantime or
skipping disks that were hot-unplugged in the meantime.

* src/qemu/qemu_domain.c (qemuDomainSnapshotForEachQcow2): Use
snapshot domain details when available.  Avoid NULL deref.

Qemu driver tries to update balloon data in virDomainGetInfo and if it
can't do so because there is another monitor job running, it just
reports what's known in domain def. However, if there was no job running
but getting the data from qemu fails, we would fail the whole API. This
doesn't make sense. Let's make the failure nonfatal.

This resolves:

     https://bugzilla.redhat.com/show_bug.cgi?id=730909

When support for setting the qemu disk error policy to "enospc" was
added, it was inadvertently spelled "enospace". This patch corrects
that on the qemu commandline (while retaining the "enospace" spelling
for libvirt's XML).

Also, while examining the qemu source, I found that "enospc" is not
allowed for the read error policy, only for write error policy (makes
sense). Since libvirt currently only has a single error policy
setting, when "enospace" is selected, the read error policy is set to
"ignore".

Destination libvirtd remembers the original name in the prepare phase
and clears it in the finish phase. The original name is used when
comparing domain name in migration cookie.

When support for was added for PCI multifunction cards (in commit
9f8baf, first included in libvirt 0.9.3), it was done by always
turning on the multifunction bit for all PCI devices. Since that time
it has been realized that this is not an ideal solution, and that the
multifunction bit must be selectively turned on. For example, see

  https://bugzilla.redhat.com/show_bug.cgi?id=728174

and the discussion before and after

  https://www.redhat.com/archives/libvir-list/2011-September/msg01036.html

This patch modifies multifunction support so that the multifunction=on
option is only added to the qemu commandline for a device if its PCI
<address> definition has the attribute "multifunction='on'", e.g.:

  <address type='pci' domain='0x0000' bus='0x00'
           slot='0x04' function='0x0' multifunction='on'/>

In practice, the multifunction bit should only be turned on if
function='0' AND other functions will be used in the same slot - it
usually isn't needed for functions 1-7 (although there are apparently
some exceptions, e.g. the Intel X53 according to the QEMU source
code), and should never be set if only function 0 will be used in the
slot. The test cases have been changed accordingly to illustrate.

With this patch in place, if a user attempts to assign multiple
functions in a slot without setting the multifunction bit for function
0, libvirt will issue an error when the domain is defined, and the
define operation will fail. In the future, we may decide to detect
this situation and automatically add multifunction=on to avoid the
error; even then it will still be useful to have a manual method of
turning on multifunction since, as stated above, there are some
devices that excpect it to be turned on for all functions in a slot.

A side effect of this patch is that attempts to use the same PCI
address for two different devices will now log an error (previously
this would cause the domain define operation to fail, but there would
be no log message generated). Because the function doing this log was
almost completely rewritten, I didn't think it worthwhile to make a
separate patch for that fix (the entire patch would immediately be
obsoleted).

Currently, qemuDomainGetXMLDesc and qemudDomainGetInfo check for
outstanding synchronous job before (eventual) monitor entering.
However, there can be already async job set, e.g. migration.