There's no need to keep @binary around.
virQEMUCapsInitGuestFromBinary() duplicates the string anyway.

 1,002 bytes in 36 blocks are definitely lost in loss record 54 of 59
    at 0x483579F: malloc (vg_replace_malloc.c:299)
    by 0x796B1C7: vasprintf (vasprintf.c:73)
    by 0x4C3F2C6: virVasprintfInternal (virstring.c:740)
    by 0x4C3F3DC: virAsprintfInternal (virstring.c:761)
    by 0x13AFC9: testGetCaps (qemucaps2xmltest.c:105)
    by 0x13B200: testQemuCapsXML (qemucaps2xmltest.c:157)
    by 0x13B642: virTestRun (testutils.c:174)
    by 0x13B366: doCapsTest (qemucaps2xmltest.c:191)
    by 0x13FF2B: testQemuCapsIterate (testutilsqemu.c:941)
    by 0x13B427: mymain (qemucaps2xmltest.c:215)
    by 0x13D706: virTestMain (testutils.c:1096)
    by 0x13B489: main (qemucaps2xmltest.c:221)
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

This function is not used anymore. Let's remove it.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NMartin Kletzander <mkletzan@redhat.com>

It's funny how this went unnoticed for such a long time. Long
story short, if a domain is configured with
VIR_DOMAIN_NUMATUNE_MEM_STRICT libvirt doesn't really honour
that. This is because of 7e72ac78 after which libvirt allowed
qemu to allocate memory just anywhere and only after that it used
some magic involving cpuset.memory_migrate and cpuset.mems to
move the memory to desired NUMA nodes. This was done in order to
work around some KVM bug where KVM would fail if there wasn't a
DMA zone available on the NUMA node. Well, while the work around
might stopped libvirt tickling the KVM bug it also caused a bug
on libvirt side: if there is not enough memory on configured NUMA
node(s) then any attempt to start a domain must fail. Because of
the way we play with guest memory domains can start just happily.

The solution is to move the child we've just forked into emulator
cgroup, set up cpuset.mems and exec() qemu only after that.

This basically reverts 7e72ac78 which was a workaround
for kernel bug. This bug was apparently fixed because I've tested
this successfully with recent kernel.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NMartin Kletzander <mkletzan@redhat.com>

This reverts commit f1d65853.

Turns out, this caused a regression. There is this (perhaps less
known) semantic of virDomainAttachDevice() where if the device
the API is trying to attach is a CDROM/floppy that is already in
the domain the attach request is handled as 'change the media in
the drive'.

We have a better fix anyways.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Tested-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NJim Fehlig <jfehlig@suse.com>

This tries to fix the same problem as f1d65853 but it's doing
so in a less invasive way.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Tested-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NJim Fehlig <jfehlig@suse.com>

Unit number 7 is kind of special. It's reserved for SCSI
controller. The comment in virDomainSCSIDriveAddressIsUsed()
summarizes that pretty nicely. Libvirt would never generate
such address.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Tested-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NJim Fehlig <jfehlig@suse.com>

This function checks if given drive address is already present in
passed domain definition. Expose the function as it will be used
shortly.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Tested-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NJim Fehlig <jfehlig@suse.com>

During initial NIC setup the hypervisor drivers are responsible for
attaching the TAP device to the bridge device. Any fixup after libvirtd
restarts should thus also be their responsibility.
Reviewed-by: NLaine Stump <laine@laine.org>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The call to resolve the actual network type will turn any NICs with
type=network into one of the other types. Thus there should be no need
to handle type=network in later switch() statements jumping off the
actual type.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Ports allocated on virtual networks with type=nat|route|open all get
given an actual type of 'network'.

Only ports in networks with type=bridge use an actual type of 'bridge'.

This distinction makes little sense since the virtualization drivers
will treat both actual types in exactly the same way, as they're all
just bridge devices a VM needs to be connected to.

This doesn't affect user visible XML since the "actual" device XML
is internal only, but we need code to convert the data upgrades.
Reviewed-by: NLaine Stump <laine@laine.org>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The virNetDevBandwidthParse method uses the interface type to decide
whether to allow use of the "floor" parameter. Using the interface
type is not convenient as callers may not have that available, but
still wish to allow use of "floor". Switch to an explicit boolean
to control its usage.
Reviewed-by: NLaine Stump <laine@laine.org>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Reword error messages to make it clear that the combined floor settings
of all NICs are exceeding the network inbound peak/average
settings. Including the actual values being checked helps to diagnose
what is actually wrong.
Reviewed-by: NLaine Stump <laine@laine.org>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

In extreme cases libvirt can get mixed up about what VMs are running and
attached to a network leading to the cached floor sum value being
outdated. When this happens the only option is to destroy the network
and then restart libvirtd. If we set floor sum back to zero when
starting the network, we avoid the need for a libvirtd restart at least.
Reviewed-by: NLaine Stump <laine@laine.org>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Replaced all virSaveLastError and virSetError/virFreeError usages to
virErrorPreserveLast and virErrorRestore respectively.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
Signed-off-by: NSyed Humaid <syedhumaidbinharoon@gmail.com>

The networkPlugBandwidth & networkUnplugBandwidth methods currently take
a virDomainNetDefPtr. To remove the dependency on the domain config
struct, pass individual parameters instead.
Reviewed-by: NLaine Stump <laine@laine.org>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

All but one of the network types supports port profiles. Rather than
duplicating the code to merge profiles 3 times, do it once and then
later report an error if used from the wrong place.
Reviewed-by: NLaine Stump <laine@laine.org>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Reviewed-by: NLaine Stump <laine@laine.org>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Firstly, VIR_STRDUP() accepts NULL, so there is no need to check
if the string we want to duplicate is not-NULL. Secondly,
virDomainNetSetModelString() also accepts NULL. Thirdly, we have
VIR_AUTOFREE().
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

It has made Clang very unhappy ever since 6bf7c676.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>

Fix for commit ffc0fbeb refactoring snapshot code.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
ACKed-by: NMaxim Nestratov <mnestratov@virtuozzo.com>
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

Fix for commit 1ab05da2 refactoring snapshot code.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
ACKed-by: NMaxim Nestratov <mnestratov@virtuozzo.com>
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

f1056279 removed virDomainSnapshotDef.current and leaved
using vm->current_snapshot only. Later 4819f54b moved current snapshot
tracking to virDomainSnapshotObjList. As vz driver never used
vm->current_snaphot this patch includes fixes after both commits.
Reviewed-by: NCole Robinson <crobinso@redhat.com>
ACKed-by: NMaxim Nestratov <mnestratov@virtuozzo.com>
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

After the recent changes, there are only a few places left
where we use the explicit path instead of taking advantage of
the publicly available define; let's get rid of those too.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

As evidenced by all existing callers, the only directory it makes
sense to use is TEST_QEMU_CAPS_PATH, so let's just bake that into
the function.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

As evidenced by all existing callers, the only directory it makes
sense to use is TEST_QEMU_CAPS_PATH, so let's just bake that into
the function.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

The value (with a slightly different name) is currently private
to testutilsqemu, but since we use this path all over the place
it makes sense to define it publicly and avoid repetition.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

Now that we can override the post-parse handling, let's update the
testsuite to provide the desired timestamp/name rather than ignoring
the non-deterministic one that was previously being generated. A few
output files need timestamps added now that they are no longer
ignored.
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

Wire up the accessor functions necessary for the testsuite to install
an alternative post-parse handler from normal drivers. I could have
modified the signature for virDomainXMLOptionNew() to take another
parameter, but thought it was easier to add a new set function rather
than chase down all existing callers. Until code actually sets the
override, there is no change in behavior.
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

Move the non-deterministic code that sets snapshot properties
independently of what the incoming XML described to instead live in a
default post-parse function common to virDomainMoment (as checkpoints
will also reuse it in later patches). This patch is just code motion,
with no difference to any callers; but the next patch will further
refactor things to allow for a per-driver override, used by the
testsuite to perform deterministic post-parse actions for better
coverage of parser/formatter code.

Note that the post-parse code is intentionally not run during a
snapshot redefine, since that code path already requires a valid
snapshot name and creation time from the XML.
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

None of the existing drivers actually use the 0-valued 'nostate'
snapshot state; rather, it was a fluke of implementation. In fact,
some drivers, like qemu, actively reject 'nostate' as invalid during a
snapshot redefine. Normally, a driver computes the state post-parse
from the current domain, and thus virDomainSnapshotGetXMLDesc() will
never expose the state. However, since the testsuite lacks any
associated domain to copy state from, and lacks post-parse processing
that normal drivers have, the testsuite output had several spots with
the state, coupled with a regex filter to ignore the oddity.

It is better to follow the lead of other XML defaults, by not
outputting anything during format if post-parse defaults have not been
applied, and rejecting the default value during parsing. The testsuite
needs a bit of an update, by adding another flag for when to simulate
a post-parse action of setting a snapshot state, but none of the
drivers are impacted other than rejecting XML that was previously
already suspicious in nature.

Similarly, don't expose creation time 0 (for now, only possible if a
user redefined a snapshot to claim creation at the Epoch, but also
happens once setting the creation time is deferred to a post-parse
handler).

This is also a step towards cleaning up snapshot_conf.c to separate
its existing post-parse work (namely, setting the creationTime and
default snapshot name) from the pure parsing work, so that we can get
rid of the testsuite hack of regex filtering of the XML and instead
have more accurate testing of our parser/formatter code.
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

Upcoming changes want to separate out a post-parse massaging of
snapshots separate from parsing the XML, so as not to be dependent on
filtering out an ever-changing timestamp from the testsuite. Along the
way, this means we will want to add yet another conditional to the
snapshot xml2xml tests on whether to perform post-processing steps to
canned values. This will be easier to read if we consolidate all the
decisions into a flags variable, instead of adding yet another
boolean.

While at it, drop the redundant inout test of "noparent" (once is
enough).
Signed-off-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NCole Robinson <crobinso@redhat.com>

This requires drivers to opt in to handle the raw modelstr
network model, all others will error if a passed in XML value
is not in the model enum.

Enable this feature for libxl/xen/xm and qemu drivers
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

Convert the vbox driver to net model enum, which requires adding
enum values for Am79C970A, Am79C973, 82540EM, 82545EM, 82543GC. We
preserve the same casing that vbox historically used for these model
names.

Remove the now unused virDomainNetStrcaseeqModelString
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

Convert the vmware/vmx driver to net model enum, which requires
adding enum values for vlance, vmxnet, vmxnet2, and vmxnet3.
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

vbox and vmx drivers do net case insensitive net model comparisons,
so for example 'VMXNET3' and 'vmxnet3' and 'VmxNeT3' in the XML will
translate to the same driver configuration. To convert these drivers
to use net model enum, we will need to do case insensitive comparisons
as well.

Essentially we implement virEnumToString, but with case insensitive
comparison. XML will always be formatted with the enum model string
we track internally, but we will accept any case insensitive variant.
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

This converts the qemu driver to the net model enum, for all
the model values that we have hardcoded for various checks,
which adds e1000e, virtio-transitional, virtio-non-transitional,
usb-net, spapr-vlan, lan9118, smc91c111

Because the qemu driver has historically also allowed the raw
model string onto the qemu command line, this isn't a full
conversion. Unwinding that will require more thought. However
for all new driver code we should be adding explicit enum
values for any model name we have special handling for.

Remove the now unused virDomainNetStreqModelString
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

The bhyve driver only works with the virtio and e1000 models,
which we already have in the enum. Some error reporting is
slightly downgraded to avoid some subtle usage of modelstr
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

The vz driver only handles three models: virtio, e1000, and rtl8139.
Add enum values for those models, and convert the vz driver to
handling net->model natively
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

This adds a network model enum. The virDomainNetDef property
is named 'model' like most other devices.

When the XML parser or a driver calls NetSetModelString, if
the passed string is in the enum, we will set net->model,
otherwise we copy the string into net->modelstr

Add a single example for the 'netfront' xen model, and wire
that up, just to verify it's all working
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

We will be adding a 'model' enum in upcoming patches. Rename
the existing value to make the differentiation clear
Acked-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>