28 bytes in 1 blocks are definitely lost in loss record 26 of 66
   at 0x4C2CF0F: malloc (vg_replace_malloc.c:299)
   by 0x7A02719: strdup (strdup.c:42)
   by 0x197DC1: virStrdup (virstring.c:961)
   by 0x12B478: virLockDaemonConfigFilePath (lock_daemon_config.c:44)
   by 0x12A759: main (lock_daemon.c:1270)

62 (32 direct, 30 indirect) bytes in 1 blocks are definitely lost in loss record 41 of 66
   at 0x4C2EF26: calloc (vg_replace_malloc.c:711)
   by 0x151B61: virAlloc (viralloc.c:144)
   by 0x12B56C: virLockDaemonConfigNew (lock_daemon_config.c:71)
   by 0x12A491: main (lock_daemon.c:1262)

13 bytes in 1 blocks are definitely lost in loss record 21 of 70
   at 0x4C2CF0F: malloc (vg_replace_malloc.c:299)
   by 0x7A02719: strdup (strdup.c:42)
   by 0x197E3F: virStrdup (virstring.c:961)
   by 0x12C86B: virLockSpaceProtocolDispatchRegister (lock_daemon_dispatch.c:291)
   by 0x12BB73: virLockSpaceProtocolDispatchRegisterHelper (lock_daemon_dispatch_stubs.h:152)
   by 0x1336AA: virNetServerProgramDispatchCall (virnetserverprogram.c:437)
   by 0x13320D: virNetServerProgramDispatch (virnetserverprogram.c:304)
   by 0x139E3E: virNetServerProcessMsg (virnetserver.c:144)
   by 0x13A1A2: virNetServerDispatchNewMessage (virnetserver.c:230)
   by 0x1350F5: virNetServerClientDispatchMessage (virnetserverclient.c:343)
   by 0x137680: virNetServerClientDispatchEvent (virnetserverclient.c:1498)
   by 0x147704: virNetSocketEventHandle (virnetsocket.c:2140)
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

The virSecurityStackAddNested() can fail in which case
virSecurityManagerNewStack() should fail too.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Use 'error' label to free allocated memory.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

The current socket test is rather crazy in that it sets up a server
listening for sockets and then runs a client connect call, relying on
the fact that the kernel will accept this despite the application
not having called accept() yet. It then closes the client socket and
calls accept() on the server. On Linux accept() will always see that
the client has gone and so skip the rest of the code. On FreeBSD,
however, the accept sometimes succeeds, causing us to then go into
code that attempts to read and write to the client which will fail
aborting the test. The accept() never succeeds on FreeBSD guests
with a single CPU, but as you add more CPUs, accept() becomes more and
more likely to succeed, giving a 100% failure rate for the test when
using 8 CPUs.

This completely rewrites the test so that it is avoids this designed in
race condition. We simply spawn a background thread to act as the
client, which will read a byte from the server and write it back again.
The main thread can now properly listen and accept the client in a
synchronous manner avoiding any races.
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The test code for UNIX and TCP sockets will need to be rewritten and
extended later, and will benefit from code sharing.
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

If no JSON parser is available qemublocktest fails, so skip its execution.
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The expected output strings from the vshtabletest.c are created on a
modern Linux host where unicode printing support is very good. On older
Linux platforms, or non-Linux platforms, some unicode characters will
not be considered printable. While the vsh table alignment code will
stil do the right thing with escaping & aligning in this case, the
result will not match the test's expected output.

Since we know the code is working correctly, do a check with iswprint()
to validate the platform's quality and skip the test if it fails. This
fixes the test on FreeBSD platforms.
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The corresponding structs are declared there.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

These are simple predicates, which makes bool a more
appropriate return type than int.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

The function is called on a virDomainDeviceInfo, so it
should be declared along with it.

Moving this function requires moving and making public
virDomainDeviceCCWAddressIsValid() as well, but that's
perfectly fine since the same reasoning above also
applies to it, due to virDomainDeviceCCWAddress being
(correctly) declared in device_conf.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

It's used in virDomainDeviceInfo, which makes
domain_conf the wrong place to declare it.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

When switching the host architecture to something for which we do not
have any host CPU model defined, the mocked
virQEMUCapsProbeHostCPUForEmulator would just return the previous CPU
model resulting in strange combinations, such as "core2duo" host CPU
model in QEMU capabilities for "AArch64" architecture. It currently
doesn't break any test case, but we should fix it anyway to avoid future
surprises which would be quite hard to debug.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1621910

When users want to update a path to a CDROM they tend to
construct a very minimal XML and feed the API with it. This is
not a good practice as it breaks the assumptions the API is built
on. Most notably, leaving an element out should be treated as a
request for removal of the corresponding setting. Just like
leaving out <bandwidth/> clears out any QoS previously set.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>

Test that we correctly accept 64-bit unsigned numbers for QEMU.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

We expect to get numbers as big as ULLONG_MAX from QEMU,
add a test for them.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Instead of printing the whole JSON in error messages,
print just the test name.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

This test gets its JSON docs from files.

Now that we have a 'name' field in testInfo, use it instead
of abusing the 'doc' field.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Give the testing function access to the test name instead of only
passing it to virTestRun.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Commit f7b55665 added 'save_error' even though the function
already has 'originalError' used in the 'try_remove' section.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

As advertised in previous commit, there are three APIs that might
run for quite some time (because they read/write data from/to a
volume) and these three are: downloadVol, uploadVol, wipeVol.
Release pool object lock and reacquire it later to allow more
concurrency.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

There are few operations in the storage driver that read/write
data onto volumes. Such operations can take very long time to
finish. During that time the storage pool object is locked which
has bad performance impacts (other threads can't fetch its XML
for instance). This commit prepares the storage driver for
releasing the lock during those operations (downloadVol,
uploadVol, wipeVol).
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Just like a few commits earlier, checking for pool source
duplicates and unlocking pools list afterwards is a buggy
pattern. The check must go into virStoragePoolObjAssignDef.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

The @conn argument is needed only to do some source matching in
case of iSCSI source. Anyway, it's used just for node device
driver and as such can be replaced with virGetConnectNodeDev().

At the same time, the @conn struct member is dropped from
_virStoragePoolObjFindDuplicateData.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

This function is going to be made static in used in
virStoragePoolObjAssignDef(). Therefore move it and all the
static functions it calls a few lines up.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Even though we do some checking it is not as thorough as it
should be. We already have virStoragePoolObjIsDuplicate but the
way we use it is a typical TOCTOU. Imagine two threads trying to
define two pools with the same name but different UUIDs. With the
current code neither of them finds a duplicate and thus proceed
to virStoragePoolObjAssignDef where only names are compared.
Therefore both threads succeed which is obviously wrong.

We should check for duplicates where we care for them.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

This function is going to be made static in used in
virStoragePoolObjAssignDef(). Therefore move it a few lines up.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

In two places the passed pool object argument is marked as
ATTRIBUTE_UNUSED even though it's used right away.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1623157

Changing MTU on a running guest is not possible and trying to do
so made us face many problems. That's why we forbid it in
5f44d7e3. However, there is still one possible path where
users can sneak in change: migration XML.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1621910

When introducing this check back in 4ad54a41 my mindset was
that if an element is missing in update XML then user is
requesting for removal of the corresponding setting. For
instance, if <bandwidth/> is not present in update XML any QoS
previously set on <interface/> is cleared out. Well this
assumption is correct but only to some extent.

Turns out, we have some users who when updating path to ISO
image construct very minimalistic disk XML and pass it to device
update API. Such XML is lacking a lot of information, and alias
is one of them. This triggers error in
virDomainDefCompatibleDevice() because we think that user is
requesting to remove the alias. Well, they are not.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

The size/capacity stats gathered in qemuDomainBlocksStatsGather when
using -blockdev would be overwritten by assigning/copying the transfered
data statistics at the end. Fix it by moving the assignment prior to
fetching the capacity data.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Signed-off-by: NPeter Krempa <pkrempa@redhat.com>

* docs/news.xml: updated for release
Signed-off-by: NDaniel Veillard <veillard@redhat.com>

Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NMartin Kletzander <mkletzan@redhat.com>

The helper function virshSnapshotCreate (formerly vshSnapshotCreate)
has had dead variables since commit a00c37f2 (Sep 2011).
Signed-off-by: NEric Blake <eblake@redhat.com>

S390 is aware of both vfio-pci and vfio-ccw devices, so
on S390 the capability QEMU_CAPS_VFIO_PCI_DISPLAY will be
available. Add an extra check to make sure we only set the
display to off for vfio-pci mediated devices. Otherwise we
add display for vfio-ccw device and this breaks vfio-ccw
device qemu command line.

Fixes: d54e45b6 conf: Introduce new <hostdev> attribute 'display'
Signed-off-by: NFarhan Ali <alifm@linux.ibm.com>
Reviewed-by: NMarc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>

The default disk storage pool type in XML is 'dos', not 'msdos'.
But tweak wording to keep the term 'msdos' in the text for the
sake of grep searches.
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Commit 6534b3c4 tried to raise an error when there is no numa
nodes by setting access='shared' in the domain config, but added
a helper called from qemuDomainDeviceDefValidate instead of a
helper called from qemuDomainDefValidate for XML:

  <memoryBacking>
    <hugepages/>
    <access mode='shared'/>
  </memoryBacking>

Since there are no memory devices in the test XML, there would
be no validation failure, but the test added was still failing.
Investigating that it turns out that unnecessary XML elements
were causing the failure (no need for <video>, <graphics>,
<pm>, usb controller model "piix3-uhci", disk attribute for
"discard='unmap'", <serial>, <console>, <channel> and a
memballoon model). Removing all those before moving the method
caused the test to succeed.

So this patch moves the validation to the right place and
removes all the unnecessary XML pieces that were causing
a false validation failure.

https://bugzilla.redhat.com/show_bug.cgi?id=1448149#c14Signed-off-by: NLuyao Huang <lhuang@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

The reason of broken build was that centos and rhel use older version of
glibc. These versions of glibc on these platforms cannot work with newer
unicodes, thus causing functions iswprint() and wcwidth() return
unexpected values causing the vshtabletest to fail. Therefore, let's
replace the new unicode characters causing issues with some older ones
to fix the test suite, as the issue would still persist during runtime.
Signed-off-by: NSimon Kobyda <skobyda@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>

If @vm has flagged as "to be removed" virDomainObjListFindByNameLocked
returns NULL (although the definition actually exists). Therefore, the
possibility exits that "virHashAddEntry" will raise the error
"Duplicate key" => virDomainObjListAddObjLocked fails =>
virDomainObjEndAPI(&vm) is called and this leads to a freeing of @def
since @def is already assigned to vm->def. But actually this leads to
a double free since the common usage pattern is that the caller of
virDomainObjListAdd(Locked) is responsible for freeing @def in case of
an error.

Let's fix this by setting vm->def to NULL in case of an error.

Backtrace:

   ➤  bt
   #0  virFree (ptrptr=0x7575757575757575)
   #1  0x000003ffb5b25b3e in virDomainResourceDefFree
   #2  0x000003ffb5b37c34 in virDomainDefFree
   #3  0x000003ff9123f734 in qemuDomainDefineXMLFlags
   #4  0x000003ff9123f7f4 in qemuDomainDefineXML
   #5  0x000003ffb5cd2c84 in virDomainDefineXML
   #6  0x000000011745aa82 in remoteDispatchDomainDefineXML
   ...
Reviewed-by: NBjoern Walk <bwalk@linux.ibm.com>
Signed-off-by: NMarc Hartmayer <mhartmay@linux.ibm.com>

We would have used virtio for networking anyway, but it's
better to be explicit; for graphics, none of the existing
models work right now but virtio is the only one which
has a non-PCI variant, so it's as good a default as any
Spotted-by: NJán Tomko <jtomko@redhat.com>
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>