This speeds up node_device_udev driver startup 11x.

Simplify the logic

Move out the code depending on HAVE_DECL_ETHTOOL_GFEATURES.

Split out the features that we probe via various ethtool commands
and ETHTOOL_GFLAGS.

Rename struct elem to virNetDevEthtoolFeatureCmd and move it
out of the function to allow reusing it.

The new flag was not added to virCheckFlags in commit '0d1c17aa' causing
a regression where VMs were not able to start.

Rather than intermixing the ATTRIBUTE_UNUSED - use HAVE_GNUTLS_CIPHER_ENCRYPT
for the whole function instead.

Filter out some subsystems we are not interested in.

Instead of the custom out and out_unlock.

Remove ret variables and labels from functions where there is no cleanup
to be done.

Introduce a validation callback for qemu and move checking of
min_guarantee to the new callback.

Now with the proper domain config validation infrastructure the check
can be moved to a place that doesn't make domains vanish.

Since it will not be called from outside of conf we can unexport it too
if we move it to the appropriate place.

Test suite change is necessary since the error will be reported sooner
now.

To avoid duplicating all the checks when starting a fresh VM from a
possibly unchecked config, call the domain def validator.

Validation of qemu process startup requires to know whether the process
is used for a fresh VM or whether it's reloaded from a
snapshot/migration. Pass this information in via a flag rather than
calculating it from a bunch of bools.

Similarly to the domain definition validator add a device validator. The
change to the prototype of the domain validator is necessary as
virDomainDeviceInfoIterateInternal requires a non-const pointer.

It's passed to all places along with the structure.

Until now we weren't able to add checks that would reject configuration
once accepted by the parser. This patch adds a new callback and
infrastructure to add such checks. In this patch all the places where
rejecting a now-invalid configuration wouldn't be a good idea are marked
with a new parser flag.

Make it obvious that the flag is controlling RNG schema validation.

There is no cleanup to be done.

Remove unnecessary ret variable and return early if we have no media
to save on indentation.

Use virTrimSpaces instead of a custom implementation.

Also use the more common "Unable to initialize mutex" string
and virReportSystemError instead of virStrerror.

They are no longer used.

Open code the call to udev_device_get_sysattr_value
in the one place where it's needed.

Callers only check for an error or a specific integer value.

The callers only care for an error, and a missing attribute
is simply NULL.

Most of the code paths free it right after converting it to
an integer.

We only care about the failure, not a missing property.

There is no need to differentiate between PROPERTY_FOUND
and PROPERTY_MISSING - we can just look if the string is non-NULL.

Historically, we added heads=1 to videos, but for example for qxl, we
did not reflect that on the command line.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1283207Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Add capabilities for both qxl and qxl-vga devices.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Error out on parsing errors and use a local const char pointer
instead of chained ifs to check whether we found a match.

Since the switch to VIR_STRDUP this function returns 1 on success,
but the caller treats any non-zero value as failure.

Use udevHasDeviceProperty instead of udevGetStringProperty.
We do not need to copy the string since we do not need it.

Also add braces around the if body, since the change made
syntax check complain.

Two out of three callers free it right after converting it to a number.

Also change the comment at the beginning of the function, because
the comment inside the function told me to.