If initializing test monitor in testQemuHotplugCpuPrepare()
fails, the control jumps to error label where
testQemuHotplugCpuDataFree() is called. But since the data->mon
is NULL due to aforementioned failure,
qemuMonitorTestGetMonitor() dereferences a NULL pointer leading
to a SIGSEGV.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>

It always loads the latest schema. Prepare for loading others as well.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>

We have a few cases validating that the code behaves correctly in
pre-modern hotplug era. This is controled by the 'modern' flag for the
test. Since 'cpu-add' command is now deprecated in qemu, there is a
modern replacement for it, and the test output is checked against
expected commands we can skip schema validation for the legacy command.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Only tests expected to fail should allow unused commads as the normal
tests will consume all of them.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Turns out that it's not enough to pass the qemu:///embed root to
virQEMUDriverConfigNew(), you also have to make sure the same
string is copied into the virQEMUDriver structure yourself, and
not doing so in our case resulted in the cleanup never happening
and in distcheck failing because of that.

On the other hand, actually setting config->embeddedRoot would
result in different paths being generated for each test run, which
would obviously break qemuxml2argvtest, so that's not an option
either.

This reverts commit d98cc196.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>

Now that the QEMU driver natively supports storing all its runtime
data inside an arbitrary directory, we can avoid having multiple
copies of the same logic in the test suite.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

We need this for all tests that use virHostdevManager, because
during creation of this object for unprivileged connections
like those used in the test suite we would end up writing inside
the user's home directory.

That's bad manners in general, but when running the test suite
inside a purposefully constrained environment such as the one
exposed by pbuilder, it turns into an outright test failure:

  Could not initialize HostdevManager - operation failed: Failed
  to create state dir '/nonexistent/.cache/libvirt/hostdevmgr'
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

To prevent unexpected situations where a change in code would stop
looking at some of the tested commands go unnoticed add a mechanism to
force consumption of all test items.

Since there are a few tests which would be hard to fix add also a
mechanism to opt-out of the check.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

On failure 'drive_del' is not issued.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

One of the test cases attempted to use test data meant for modern qemu
without asserting the 'modern' flag. Since that changes the commands
used to query state it won't work with data meant for the modern case.

Remove the invalid test case.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

The test verifies unplug of a disk with the persistent reservations
helper. The 'object-del' used to remove the helper was not mentioned in
the list of expected commands.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

The 'human-monitor-command' equates to the 'drive-del' command issued by
the hotplug code on successful detach of a device. This means that it's
not issued for failed attempts and thus should not be added to the
expected list. Unfortunately our test monitor doesn't ensure that all
expected commands were consumed.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

The qemuMonitorTestNew() function returns with the monitor object
locked, and expects it to still be locked when qemuMonitorTestFree
is called.  The qemuhotplug test, however, explicitly unlocks the
monitor, but then forgets to lock it again. As a result the
qemuMonitorTestFree function is unlocking a mutex it doesn't own.

This bug has existed forever, but since we use normal POSIX mutexes
and don't check the return value of pthread_mutex_lock/unlock we
didn't see the error. It was harmless until the switch to the per
monitor event loop which requires the thread synchronization to
work reliably, whereupon it started crashing.
Reviewed-by: NPeter Krempa <pkrempa@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

If a disk has persistent reservations enabled, qemu-pr-helper
might open not only /dev/mapper/control but also individual
targets of the multipath device. We are already querying for them
in CGroups, but now we have to create them in the namespace too.
This was brought up in [1].

1: https://bugzilla.redhat.com/show_bug.cgi?id=1711045#c61Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Tested-by: NLin Ma <LMa@suse.com>
Reviewed-by: NJim Fehlig <jfehlig@suse.com>

Use the new command in the test suite by asserting the capability
and adjusting test data to the correct field names as they changed
compared to 'query-cpus'.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

There are some architectures that support capabilities that others
do not (e.g. s390x supports cpu comparison and baseline via QEMU).

Let's make testQEMUSchemaLoad accept a string to specify the schema
to load based on the specified arch.
Signed-off-by: NCollin Walling <walling@linux.ibm.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

If virHostdevManagerGetDefault in qemuhotplugtest fails it works
for quite a while to later segfault when accessing
mgr->activePCIHostdevs.

Report the error details and break on a failed init to see the
real issue right away.
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>
Signed-off-by: NChristian Ehrhardt <christian.ehrhardt@canonical.com>

Eliminate direct use of normal setenv/unsetenv calls in
favour of GLib's wrapper. This eliminates two gnulib
modules
Reviewed-by: NFabiano Fidêncio <fidencio@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Move the SPICE caps validation from qemuBuildGraphicsSPICECommandLine()
to a new function called qemuDomainDeviceDefValidateSPICEGraphics().
This function is called by qemuDomainDeviceDefValidateGraphics(),
which in turn is called by qemuDomainDefValidate(), validating the graphics
parameters in domain define time.

This validation move exposed a flaw in the 'default-video-type' tests
for PPC64, AARCH64 and s390 archs. The XML was considering 'spice' as
the default video type, which isn't true for those architectures.
This was flying under the radar until now because the SPICE validation
was being made in 'virsh start' time, while the XML validation done in
qemuxml2xmltest.c considers define time.

All other tests were adapted to consider SPICE validation in this
earlier stage.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

Move the VNC cap validation from qemuBuildGraphicsVNCCommandLine()
to qemuDomainDeviceDefValidateGraphics(). This function is called by
qemuDomainDefValidate(), validating the graphics parameters in domain
define time.

Tests were adapted to consider SDL validation in this earlier stage.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

Move the PM validation being done by qemuBuildPMCommandLine() to
to a new qemuDomainDefValidatePM() function. This new function
is called by qemuDomainDefValidate(), promoting PM validation in
domain define time.

Tests were adapted to consider the new caps being needed in
this earlier stage.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

As of commit 2a00ef6e which
was released in v5.2.0, we require YAJL to build the QEMU driver.

Remove the checks from code that requires the QEMU driver
or checks that also check for WITH_QEMU.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

This parameter is now unused and can be removed entirely.
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Moving their instance parameter to be the first one, and give consistent
ordering of other parameters across all functions. Ensure that the xml
options are passed into both functions in prep for future work.
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

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

Some of our tests try to validate domain XMLs they are working
with (not intentionally, simply because they call top level
domain XML parse function). Anyway, this implies that we build
domain capabilities also - see
virQEMUDriverGetDomainCapabilities(). And since some domain XMLs
are type of 'kvm' the control gets through
virQEMUCapsFillDomainCaps() and virHostCPUGetKVMMaxVCPUs() to
opening /dev/kvm which may be missing on the machine we're
running 'make check'.

Previously, we did not see this issue, because it was masked. If
building domain capabilities failed for whatever reason, we
ignored the failure. Only v5.9.0-207-gc69e6ede uncovered the
problem (it changed reval from 0 to -1 if
virQEMUDriverGetDomainCapabilities() fails). Since the referenced
commit is correct, we need to mock access to /dev/kvm in our
tests.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Prefer the GLib version to the one from gnulib.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NPeter Krempa <pkrempa@redhat.com>

Several tests were not specifying the necessary qemu capabilities for
what they were testing. Due to the way that the video devices are
currently validated, this is not causing any problems. But a change to
video device validation in a following patch would have exposed this
issue and resulted in multiple test failures about the domain
configuration not supporting particular video models.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Signed-off-by: NJonathon Jongsma <jjongsma@redhat.com>

qemu_hotplugpriv.h is a header file created to share a global variable
called 'qemuDomainRemoveDeviceWaitTime', declared in qemu_hotplug.c,
to other files that would want to change the timeout value
(currently, only tests/qemuhotplugtest.c).

Previous patch deprecated the variable, using qemu_driver->unplugTimeout
to set the timeout instead. This means that the header file is now
unused, and can be safely discarded.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

For some architectures and setups, device removal can take
longer than the default 5 seconds. This results in commands
such as 'virsh setvcpus' to fire timeout messages even if
the operation were successful in the guest, confusing the
user.

This patch sets a new 10 seconds unplug timeout for PPC64
guests. All other archs will keep the default 5 seconds
timeout.

Instead of putting 'if PPC64' conditionals inside qemu_hotplug.c
to set the new timeout value, a new function called
qemuDomainGetUnplugTimeout was added. The timeout value is then
retrieved when needed, by passing the correspondent DomainDef
object. This approach allows for different guest architectures
to have distint unplug timeout intervals, regardless of the
host architecture. This design also makes it easier to
modify/enhance the unplug timeout logic in the future
(allow for special timeouts for TCG domains, for example).

A new mock file was created to work with qemuhotplugtest.c,
given that the test timeout is significantly shorter than
the actual timeout value in qemu_hotplug.c.

The now unused 'qemuDomainRemoveDeviceWaitTime' global can't
be simply erased from qemu_hotplug.c though. Next patch will
remove it properly.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Suggested-by: NCole Robinson <crobinso@redhat.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

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

Commit 944a35d7 added @fakerootdir; however, there are multiple
paths out of mymain that didn't free the memory - so just use the
g_autofree to resolve the potential leak.

Found by Coverity
Signed-off-by: NJohn Ferlan <jferlan@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Replace all occurrences of
  if (VIR_STRDUP(a, b) < 0)
     /* effectively dead code */
with:
  a = g_strdup(b);
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

Since commit 44e7f029
    util: rewrite auto cleanup macros to use glib's equivalent

VIR_AUTOPTR aliases to g_autoptr. Replace all of its use by the GLib
macro version.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

Neither virThreadInitialize or virThreadOnExit do anything since we
dropped the Win32 threads impl, in favour of win-pthreads with:

  commit 0240d94c
  Author: Daniel P. Berrange <berrange@redhat.com>
  Date:   Wed Jan 22 16:17:10 2014 +0000

      Remove windows thread implementation in favour of pthreads
Reviewed-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: NDaniel P. Berrangé <berrange@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>

Similarly to the previous commit, VIR_TEST_VERBOSE should put
'\n' at the end of each call so that the output is not broken.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Since qemuDomainDeviceDefPostParse callback requires qemuCaps, we need
to make sure it gets the capabilities stored in the domain's private
data if the domain is running. Passing NULL may cause QEMU capabilities
probing to be triggered in case QEMU binary changed in the meantime.
When this happens while a running domain object is locked, QMP event
delivered to the domain before QEMU capabilities probing finishes will
deadlock the event loop.

QEMU capabilities lookup (via domainPostParseDataAlloc callback) is
hidden inside virDomainDeviceDefPostParseOne with no way to pass
qemuCaps to virDomainDeviceDef* functions. This patch fixes all
remaining paths leading to virDomainDeviceDefPostParse.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>