Rewrite some parts for clarity, elaborate the meaning of some of the XML
attributes.  And where necessary, distinguish that we're dealing with
two different XML documents here:

  - the domainCapabilities XML, to detect the host "hypervisor"
    (QEMU/KVM) capabilities, and what libvirt knows about them.

  - the guest XML definition, i.e. what features a guest can use, based
    on the capabilities (of QEMU and libvirt and the host) reported in
    the domainCapabilities XML.
Signed-off-by: NKashyap Chamarthy <kchamart@redhat.com>
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

Currently the RPM spec doesn't add the 'secboot'-variant OVMF binaries
(an unintentional omission, checking with Cole on #virt, OFTC) for
'x86_64' and 'ia32'.  Add them.

This way, getDomainCapabilities() will report all the OVMF binaries that
are present on the system.  E.g. on Fedora 29, if you only have the
edk2-ovmf-20190308stable-1.fc29.noarch package installed, then running
`virsh domcapabilities` will enumerate _both_ the OVMF binaries (instead
of just the OVMF_CODE.fd):

  $> virsh getdomcapabilities
    ...
    <loader supported='yes'>
      <value>/usr/share/edk2/ovmf/OVMF_CODE.fd</value>
      <value>/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</value>
    ...

(
Learnt this from a discussion with Michal Privoznik in this bug,
comment#2:

    https://bugzilla.redhat.com/show_bug.cgi?id=1733940 -- RFE: Report
    firmware (FW) paths in domainCapabilities based on FW descriptor
    files
)
Signed-off-by: NKashyap Chamarthy <kchamart@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

The snapshot-create operation of running guests saves the live
XML and uses it to replace the active and inactive domain in
case of revert. So, the config XML is ignored by the snapshot
process. This commit changes it and adds the config XML in the
snapshot XML as the <inactiveDomain> entry.

In case of offline guest, the behavior remains the same and the
config XML is saved in the snapshot XML as <domain> entry. The
behavior of older snapshots of running guests, that don't have
the new <inactiveDomain>, remains the same too. The revert, in
this case, overrides both active and inactive domain with the
<domain> entry. So, the <inactiveDomain> in the snapshot XML is
not required to snapshot work, but it's useful to preserve the
config XML of running guests.
Signed-off-by: NMaxiwell S. Garcia <maxiwell@linux.ibm.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Tested-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NJiri Denemark <jdenemar@redhat.com>

The function virDomainDefFormatInternal() has the predefined root name
"domain" to format the XML. But to save both active and inactive domain
in the snapshot XML, the new root name "inactiveDomain" was created.
So, the new function virDomainDefFormatInternalSetRootName() allows to
choose the root name of XML. The former function became a tiny wrapper
to call the new function setting the correct parameters.
Signed-off-by: NMaxiwell S. Garcia <maxiwell@linux.ibm.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Tested-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NJiri Denemark <jdenemar@redhat.com>

Once we copy the domain definition from virDomainSnapshotDef, we either
need to assign it to the domain object or free it to avoid memory leaks.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Commit f1056279 introduced a regression: if reverting to a snapshot
fails early (such as when we refuse to revert to an external
snapshot), we lose track of the domain's current snapshot.

Before that patch, we were tracking the notion of the domain's current
snapshot via two means: vm->current_snapshot (which was left untouched
on early exit) and snap->def->current (which only controls what gets
written to XML to remember snapshots across libvirtd restarts).  That
patch was fixing a real bug: if a revert operation failed early, later
questions from the same libvirtd did not see any change to the current
snapsthot, but restarting libvirtd would now claim there is no current
snapshot.  But it fixed it in the wrong direction, in that the current
snapshot was forgotten unconditionally, rather than only when the
snapshot to revert to has a chance of being useful.

It didn't help that the code after that patch had two separate spots
clearing the old notion of the current snapshot - one after
determining the snapshot to revert to was viable, the other
unconditionally on all failure exit paths.  At any rate, the fix is
simple: drop the unconditional cleanup on error paths, and rely only
on the normal cleanup after early checks.

Sadly, it is not possible to test this bug in the existing
tests/virsh-snapshot, as the test driver does not have the same
prohibition against reverting to an external snapshot as the qemu
driver.

See: https://bugzilla.redhat.com/1738747Signed-off-by: NEric Blake <eblake@redhat.com>
Message-Id: <20190909205242.15406-1-eblake@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

In recent commit of 3d21ff72 the virNetDevMacVLanTapOpen() and
virNetDevMacVLanTapSetup() functions were exported in our private
symbols. But these functions live in an #ifdef so they need a
stub implementation.
Then in 1b46566e the virNetDevMacVLanIsMacvtap() function was
implemented but again, only for #idef and without stub.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

The Perl bindings for libvirt use the test driver for unit tests. This
tries to load the cpu_map/index.xml file, and when run from an
uninstalled build will fail.

The problem is that virFileActivateDirOverride is called by our various
binaries like libvirtd, virsh, but is not called when a 3rd party app
uses libvirt.so

To deal with this we allow the LIBVIRT_DIR_OVERRIDE=1 env variable to be
set and make virInitialize look for this. The 'run' script will set it,
so now build using this script to run against an uninstalled tree we
will correctly resolve files to the source tree.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The only caller for which this check makes sense is virDomainDefParse.
Thus the check should be moved there.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

This reverts commit 39dded7b.

This commit broke virpolkittest on Ubuntu 18 which has an old
dbus (v1.12.2). Any other distro with the recent one works
(v1.12.16) which hints its a bug in dbus somewhere. Revert the
commit to stop tickling it.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NMarc-André Lureau <marcandre.lureau@redhat.com>

Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

There are two 'cleanup' labels - one in
virQEMUDriverConfigHugeTLBFSInit() and the other in
virQEMUDriverConfigSetDefaults() that do nothing more than
return and integer value. No memory freeing or anything important
is done there. Drop them in favour of returning immediately.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Our naming rules prefer qemuObjectOperation() scheme rather than
qemuOperationObject() for function names. These were not honoured
in recent commits to qemu_conf.c.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

Traditionally, macvtap devices are supported using <interface
type='direct'>, but that type requires specifying a source device name
and macvtap mode which can't be altered after the initial device
creation (and may not even be available to the management software
that's creating the XML config to feed to libvirt).

But the attributes in the <source> are essentially describing how the
device will be connected to the network, and if libvirt is to be
supplied with the name of a macvtap device that has already been
created, that device will also already be connected to the network
(and the connection can't be changed). Thus it seems more appropriate
to use type='ethernet', which was created explicitly for this purpose
- for devices that have already been (or will be) connected to the
external network by someone/something outside of libvirt. The fact
that it is a *macv*tap rather than a contentional tap device is just a
detail.

This patch supports using an existing macvtap device with <interface
type='ethernet'> by checking the supplied target dev name to see if it
is a macvtap device and, when this is the case, calling
virNetDevMacVLanTapOpen() instead of virNetDevTapCreate(). For
consistency, this is only done when target managed='no'.

Resolves: https://bugzilla.redhat.com/1723367 (partially)
Signed-off-by: NLaine Stump <laine@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

If managed='no', then the tap device must already exist, and setting
of MAC address and online status (IFF_UP) is skipped.

NB: we still set IFF_VNET_HDR and IFF_MULTI_QUEUE as appropriate,
because those bits must be properly set in the TUNSETIFF we use to set
the tap device name of the handle we've opened - if IFF_VNET_HDR has
not been set and we set it the request will be honored even when
running libvirtd unprivileged; if IFF_MULTI_QUEUE is requested to be
different than how it was created, that will result in an error from
the kernel. This means that you don't need to pay attention to
IFF_VNET_HDR when creating the tap devices, but you *do* need to set
IFF_MULTI_QUEUE if you're going to use multiple queues for your tap
device.

NB2: /dev/vhost-net normally has permissions 600, so it can't be
opened by an unprivileged process. This would normally cause a warning
message when using a virtio net device from an unprivileged
libvirtd. I've found that setting the permissions for /dev/vhost-net
permits unprivileged libvirtd to use vhost-net for virtio devices, but
have no idea what sort of security implications that has. I haven't
changed libvrit's code to avoid *attempting* to open /dev/vhost-net -
if you are concerned about the security of opening up permissions of
/dev/vhost-net (probably a good idea at least until we ask someone who
knows about the code) then add <driver name='qemu'/> to the interface
definition and you'll avoid the warning message.

Note that virNetDevTapCreate() is the correct function to call in the
case of an existing device, because the same ioctl() that creates a
new tap device will also open an existing tap device.

Resolves: https://bugzilla.redhat.com/1723367 (partially)
Signed-off-by: NLaine Stump <laine@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Although <interface type='ethernet'> has always been able to use an
existing tap device, this is just a coincidence due to the fact that
the same ioctl is used to create a new tap device or get a handle to
an existing device.

Even then, once we have the handle to the device, we still insist on
doing extra setup to it (setting the MAC address and IFF_UP).  That
*might* be okay if libvirtd is running as a privileged process, but if
libvirtd is running as an unprivileged user, those attempted
modifications to the tap device will fail (yes, even if the tap is set
to be owned by the user running libvirtd). We could avoid this if we
knew that the device already existed, but as stated above, an existing
device and new device are both accessed in the same manner, and
anyway, we need to preserve existing behavior for those who are
already using pre-existing devices with privileged libvirtd (and
allowing/expecting libvirt to configure the pre-existing device).

In order to cleanly support the idea of using a pre-existing and
pre-configured tap device, this patch introduces a new optional
attribute "managed" for the interface <target> element. This
attribute is only valid for <interface type='ethernet'> (since all
other interface types have mandatory config that doesn't apply in the
case where we expect the tap device to be setup before we
get it). The syntax would look something like this:

   <interface type='ethernet'>
      <target dev='mytap0' managed='no'/>
      ...
   </interface>

This patch just adds managed to the grammar and parser for <target>,
but has no functionality behind it.

(NB: when managed='no' (the default when not specified is 'yes'), the
target dev is always a name explicitly provided, so we don't
auto-remove it from the config just because it starts with "vnet"
(VIR_NET_GENERATED_TAP_PREFIX); this makes it possible to use the
same pattern of names that libvirt itself uses when it automatically
creates the tap devices.)
Signed-off-by: NLaine Stump <laine@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

This will simplify addition of another attribute to the <target> element
Signed-off-by: NLaine Stump <laine@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

This just moves around a few things in qemuInterfaceConnect() with no
functional difference (except that a few failures that would have
previously resulted in a "success" audit log will now properly produce
a "fail" audit). The change is so that adding support for unmanaged
tap/macvtap devices will be more easily reviewable.
Signed-off-by: NLaine Stump <laine@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

In virNetDevMacVLanOpen(), The "retries" arg has been removed and the
value hardcoded as 10, since previously the function was only called
from one place, so it was always 10.
Signed-off-by: NLaine Stump <laine@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

This function returns T if the given name is a macvtap device. This is
determined by 1) getting the ifindex of the device with that name (if
there is one), and 2) checking for existence of /dev/tapXX, where "XX"
is the ifindex learned in (1).

It's also possible to learn this by getting a netlink dump of the
interface and parsing through it to look for some attributes, but that
is complicated to figure out, takes longer to execute, and I'm lazy.
Signed-off-by: NLaine Stump <laine@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

There are already good number of test cases with hostdevices,
few have multifunction devices but none having more than one
than one multifunction cards.

This patch adds a case where there are two multifunction cards
and two Virtual functions part of the same XML.

0001:01:00.X & 0005:09:00.X - are Multifunction PCI cards.
0000:06:12.[5|6] - are SRIOV Virtual functions

Future commits will improve on automatically detecting the
multifunction cards and auto-assinging the addresses
appropriately.
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NShivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

Previous patch had to add '/sys/kernel/' prefix in opendir() because
the path, which is being mocked, wasn't being considered due to
an 'if SYSFS_PCI_PREFIX' guarding the call to getrealpath().

In fact, all current getrealpath() callers are guarding it with a
conditional to ensure that the function will never be called with
a non-mocked path. In this case, an extra non-NULL verification is
needed for the 'newpath' string to use the variable - which is
counterintuitive, given that getrealpath() will always write the
'newpath' string in any non-error conditon.

However, simply removing the guard of all getrealpath() instances
causes an abort in init_env(). This happens because tests will
execute access() to non-mocked paths even before the
LIBVIRT_FAKE_ROOT_DIR variable is declared in the test files. We
don't need 'fakerootdir' to be created at this point though.

This patch does the following changes to simplify getrealpath()
usage:

- getrealpath() will now guard the init_env() call by checking if
both fakeroot isn't created and the required path is being mocked.
This ensures that we're not failing inside init_env() because
we're too early and LIBVIRT_FAKE_ROOT_DIR wasn't defined yet;

- remove all conditional guards to call getrealpath() from
access(), virMockStatRedirect(), open(), open_2(), opendir()
and virFileCanonicalizePath(). As a bonus, remove all ternary
conditionals with 'newpath';

- a new 'pathPrefixIsMocked()' helper to aggregate all the prefixes
we're mocking, making it easier to add/remove them. If a prefix
is added inside this function, we can be sure that all functions
are mocking them.
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

This patch adds hostdev test cases in qemuhotplugtest.c.

Note: the small tweak inside virpcimock.c was needed because
the new tests added a code path in which virHostHasIOMMU()
(virutil.c) started being called, and the mocked '/sys/kernel/'
prefix that is mocked in virpcimock.c wasn't being considered
in the opendir() mock. An alternative to avoid these situations
in virpcimock.c is implemented in the next patch.
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NShivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

The softlink to physfn is the way to know if the device is
VF or not. So, the patch softlinks 'physfn' to the parent function.
The multifunction PCI devices dont have 'physfn' softlinks.

The patch adds few Virtual functions to the mock environment and
changes the existing VFIO test xmls using the VFs to use the newly
added VFs for their use case.
Signed-off-by: NShivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

This patch adds mock of the /dev/vfio path, needed for proper
implementation of the support for multifunction/multiple devices
per iommu groups.

To do that, the existing bind and unbind operations were adapted
to operate with the mocked filesystem as well.
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

When starting a domain, we use the presence of a vfio-pci or
mdev hostdev to determine if the memlock maximum needs to be
increased.  But if we hotplug either of these devices, only the
vfio-pci path gets that love.  This means that attaching a, say,
vfio-ccw device will appear to succeed but the device may be
unusable as the guest may see I/O errors on long CCW chains.
The host, meanwhile, would be flooded with these messages:

  vfio_pin_page_external: Task qemu-system-s39 (11584) RLIMIT_MEMLOCK (65536) exceeded

Let's adjust the maximum memlock value in the mdev hotplug path,
so that the domain has the same value as if it were started with
one or more mdev devices in its configuration.
Signed-off-by: NEric Farman <farman@linux.ibm.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

If attaching a PCI hostdev fails, there are several things that
need to be un-done as part of the cleanup.  One thing that is
not done is re-calculating/re-setting the maximum amount of locked
memory for the domain, since we may have changed that.

Let's fix that, just to ensure everything is back the way it was.
Signed-off-by: NEric Farman <farman@linux.ibm.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Let's pull this hunk out into a function, so it can be reused
in another codepath that needs to do the same thing.
Signed-off-by: NEric Farman <farman@linux.ibm.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

In f08e6883 I've made @pcidevs in
virHostdevReAttachPCIDevices() to be automatically unrefed using
VIR_AUTOUNREF() but I forgot to remove the line that explicitly
unrefs the object at the end of the function.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

We currently generate two completely separate API references for the
libvirt public API. One at 'docs/html/' and one at 'docs/devhelp/'.
Both are published on the website, but we only link to content in
the 'docs/html/' pages.

Both are installed in the libvirt-docs sub-RPM, with a full copy
of the website including 'docs/html/' in /usr/share/docs/libvirt-docs,
while the 'docs/devhelp/' content goes to /usr/share/gtk-doc/. The
latter was broken for years until:

  commit ca6f6025
  Author: Andrea Bolognani <abologna@redhat.com>
  Date:   Fri May 10 14:54:52 2019 +0200

    docs: Introduce $(devhelphtml_generated)

    Our XSLT magic generates one Devhelp-compatible HTML file
    per documentation module, but so far we have only shipped
    and installed documentation for virterror.

    Now that we have $(modules), however, we can generate the
    list of files the same way we do for regular documentation
    and make sure we always ship and install everything.

That this bug went unnoticed for so long is a sign of how few
people are using the devhelp docs. The only commits to the devhelp
code since it was first introduced have been fixing various build
problems that hit.

The only obvious difference between the two sets of docs is the CSS
styling in use. Overall devhelp does not look compelling enough to
justify having two duplicated sets of API docs. Eliminating it will
reduce the amount of XSL code we are carrying in the tree which is
an attractive benefit.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

After the previous commits, qemuAddSharedDevice() and
qemuRemoveSharedDevice() are now the same code with a different
flag to call the internal functions.

This patch aggregates the common code into a new function called
qemuAddRemoveSharedDeviceInternal() to further reduce
code repetition. Both qemuAddSharedDevice() and
qemuRemoveSharedDevice() are kept since they are public
functions used elsewhere.

No functional change was made.
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Following the same idea of avoid code repetition from the
previous patch, this commit introduces a new function that
aggregates the functions of qemuAddSharedDisk() and
qemuRemoveSharedDisk() into a single place, using a flag to
switch between add/remove operations.

Both qemuAddSharedDisk() and qemuRemoveSharedDisk() are
public, so keep them around to avoid changing other files
due to an internal qemu_conf.c refactory.

No functional change was made.
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

qemuAddSharedHostdev() has a code similar to
qemuRemoveSharedHostdev(), with exception of one line that
defines the operation (add or remove).

This patch introduces a new function that aggregates the common
code, using a flag to switch between the operations, avoiding
code repetition.

No functional change was made.
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Since commit fd9ef3b3, virDomainFindByUUIDRef() no longer exists and
all virDomainObjListFindBy*() functions now increment the reference
count.
Signed-off-by: NJonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

As a result of changes in

      commit d5f0c1b6
      Author: Daniel P. Berrangé <berrange@redhat.com>
      Date:   Thu Jul 18 12:30:22 2019 +0100

        remote: stop trying to print help as giant blocks of text

The socket path built would be libvirt//var/run/libvirt-sock
instead of /var/run/libvirt/libvirt-sock. Fortunately this only
affects users who have set the 'unix_sock_dir' config parameter
in /etc/libvirt/libvirtd.conf, which is pretty rare/unusual.
Signed-off-by: Neater <=@eater.me>

Exception made for the psuedonym above since patch is considered
trivial & thus non-copyrightable material.
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

In some places where virDomainObjListForEach() is called the
passed callback calls virDomainObjListRemoveLocked(). Well, this
is unsafe, because the former only grabs a read lock but the
latter modifies the list.
I've identified the following unsafe calls:

- qemuProcessReconnectAll()
- libxlReconnectDomains()

The rest seem to be safe.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>