Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

qemu.control.kernel-version: Bug fix for control.kernel-version

Shared: Add new guests Win8 and Win2012

Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

Signed-off-by: NChris Evich <cevich@redhat.com>

Signed-off-by: NJiří Župka <jzupka@redhat.com>

* Fix incorrect command when run test with qmp monitor;
* Fix logical error that target image file should not
  be opened before the reopen step;
* Merge configuration file into drive_mirror.cfg
* Use by default upstream block mirror names rather than
  assuming RHEL block mirror names.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>
Signed-off-by: NXu Tian <xutian@redhat.com>

Refactor qemu_storage.QemuImg() so that it's more logical
to query for available qemu-img commands. Also, introduce
a method .compare_images() method that will take 2 images
and compare them using the subcommand 'compare', if it is
available.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

virttest.qemu_vm: Remove init_pci_addr from VM.

To add test case for virsh nodememstats command

Those files were added there mainly to mirror qemu,
since that is used as a module, since for that
case we do use things like

from qemu.tests import block_copy

That's not needed for libvirt, and we'd like to
dump it from qemu as well, although it needs
more care.

This solves a problem where it's not possible
to import the python library bindings on virt-test,
since the libvirt module present in virt-test would
take precedence over the library installed system
wide.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

QEMU-KVM is no more, and spice was split in spice common and
protocol, so let's update the build.cfg file with that
information.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Add the proper origin, and fetch submodules for git repos,
that'll make it possible to build QEMU with spice support
from source.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

* Be less pedantic about re-formatting already cached data
* Added 'url' key to standardized issue format
* Added two useful sample report scripts
Signed-off-by: NChris Evich <cevich@redhat.com>

Signed-off-by: NChris Evich <cevich@redhat.com>

Add qmp_command test case.

Signed-off-by: NYu Mingfei <yumingfei@cn.fujitsu.com>

Signed-off-by: Nyangdongsheng <yangds.fnst@cn.fujitsu.com>

changes from v1:
(1).use getfattr to replace ls -Z
(2).use /proc//attr/current to replace ps -Z
(3).remove some error type seems unnecssary.
Signed-off-by: Nyangdongsheng <yangds.fnst@cn.fujitsu.com>

Signed-off-by: Nyangdongsheng <yangds.fnst@cn.fujitsu.com>

Signed-off-by: Nyangdongsheng <yangds.fnst@cn.fujitsu.com>

(1).add a define method for dummy_virsh.
(2).some small fix to make pylint calm down.
Signed-off-by: Nyangdongsheng <yangds.fnst@cn.fujitsu.com>

Test may timeout when check guest's cpu after hotplug.
Update the timeout = 60 + total_cpus.

Update the log message when max cpu reached.
Set autotest case timeout to 2000s in add_max_cpu subcase.
run_autotest always timeout in add_max_cpu case.
Signed-off-by: NFeng Yang <fyang@redhat.com>

With the work made on the asset library, we forgot to
pass the force param to asset.download_file(), as a
result the test runner --restore-image flag stopped
working.

So fix the problem by passing the appropriate option
to that function.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

This patch just makes the cases work. I will change cases really
use multi_queue by another split patchset.
Signed-off-by: NAmos Kong <akong@redhat.com>

Multiqueue feature was supported by kernel/qemu upstream already,
this patch changed Autotest to support command line.

This change compat old qemu/kernel, it will still use 'fd' when
there is only one queue.

More information about multiqueue:
  http://www.linux-kvm.org/page/Multiqueue
  linux/Documentation/networking/tuntap.txt
Signed-off-by: NAmos Kong <akong@redhat.com>

Signed-off-by: NAmos Kong <akong@redhat.com>

Valid ifname is provided and used in open_tap(), the
evaluation in the end is redundant, just remove it.
Signed-off-by: NAmos Kong <akong@redhat.com>

List some case only Host_RHEL.6
Do not run these case in RHEL3, RHEL4 guest.
Reconfig qmp_query-mice test case.
Rename qmp_device_add to qmp_redhat_device_add.
Re-design qmp_command.qmp_query-version test case
Signed-off-by: NFeng Yang <fyang@redhat.com>

Signed-off-by: NFeng Yang <fyang@redhat.com>

This script have coverd following qmp command:
query-hpet, query-kvm, query-mice, query-status
query-name, query-uuid, query-version, query-chardev
query-blockstats, memsave, pmemsave, query-cpus
cpu, query-vnc, query-block, query-commands
query-balloon, balloon, cont, stop
and a negative test with wrong command.

This script will check result by following way:
cmd_result_check = equal, will compare cmd_return_value with qmp command output.
cmd_result_check = contain, will try to find cmd_return_value in qmp command output.
cmd_result_check = m_equal_q, will compare key value in monitor command output and qmp command output.
cmd_result_check = m_in_q, will try to find monitor command output's key value in qmp command output.
cmd_result_check = post_contain, will try to find cmd_return_value in post command output.

This update is base on upstream QMP patchset.

Changes from v1:
Bug fix and update according to ldoktor's comment
Signed-off-by: NFeng Yang <fyang@redhat.com>
KVM-test: Add format check for qmp_command test
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

This patch fix three problems in control.kernel-version:
  - Fix the wrong variable name.
  - Remove the parse_file inside control file for host-kernel.cfg
    Just in clude it as other cfg files in your tests-example.cfg
    which is more flexible. And this also can work with the situation
    that without host-kernel.cfg.
  - Don't get the package version if there is no need for package
    version check.

Also add some docstring for how to use the package filter.
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

Now vm.init_pci_addr is effected by previous vm.

We set init_pci_addr to 3 in multi_disk.max_disk, but this case still
fail for:
VMPCIOutOfRangeError: Too many PCI devices added on vm 'virt-tests-vm1',
 max supported '31'

because previous case use default value 4.

As init_pci_addr only used in get_free_pci_addr, so remove it from vm
 and only get init_pci_addr in get_free_pci_addr().
Also set default for init_pci_addr to 3.
Signed-off-by: NFeng Yang <fyang@redhat.com>

qemu: Add 3 tests steps for usb testing

* Check dmesg output after booting with usb device.
* Negative testing for usb hotplug
* Assign max number of devices (8) to usb-hub and check if qemu works well.

Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

Signed-off-by: NQingtang Zhou <qzhou@redhat.com>