This patch add test for qemu systemtap tracing
Signed-off-by: NYunping Zheng <yunzheng@redhat.com>

Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

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

Check physical block size and logical block size for virtio block device:
1) Install guest with a new image.
2) Verify whether physical/logical block size in guest is same as qemu
   parameters.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>
Acked-by: Feng Yang<fyang@redhat.com>
Acked-by: NYiqiao Pu <ypu@redhat.com>
Signed-off-by: NFeng Yang <fyang@redhat.com>

1) Create image without cluster_size option
2) Verify if the cluster_size is default value
3) Create image with cluster_size option
4) Verify if the cluster_size is the value set
Signed-off-by: NYiqiao Pu <ypu@redhat.com>
Acked-by: NQingtang Zhou <qzhou@redhat.com>

This patch using iperf test guest nic multicast, support both
linux and windows guests.
Signed-off-by: NYunping Zheng <yunzheng@redhat.com>

This patch make case jumbo test support windows except winxp
Signed-off-by: NYunping Zheng <yunzheng@redhat.com>

This patch makes mac_change case support windows guests, except WinXP
Signed-off-by: NYunping Zheng <yunzheng@redhat.com>

This patch adds function support restart windows guest network,
except winxp. add function to get windows nic attributes(like nic index,
netconnectionid, macaddress ans so on, you can get all support
attributes by 'wmic nic get ').
Signed-off-by: NYunping Zheng <yunzheng@redhat.com>

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

change from v1:
Use the new solution to resolve index problem.
It can be an sequential index generator, and also can test some special
situation that we want set some devices to specific index.

index_in_use = {} -- if there are definite index items, will mark them
                   in this dict. eg: drive_index_cd1 = 6, {'6': True}
index_global = 0 -- it is global variable, use it for image and cd's index.

if need set the specific index for stg or cd, can use following format:
drive_index_$name = $number
eg:  dirve_index_image1 = 5
     drive_index_stg2 = 6
     drive_index_cd1 = 7

change from v2:
delete global variables. index_in_use & index_global,
add member index_in_use in qemu_vm.VM class
Signed-off-by: NGolita Yue <gyue@redhat.com>

Since the defaults is generally used in config files for variants
and comments, it's a bit hard to grep the 'real' parameter.

And the meaning of 'defaults' is not much clear, so rename it to
'use_nodefaults', to mark its relationship with '-nodefaults'
option in qemu.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

monitor.migrate_set_speed only support string now
Signed-off-by: NFeng Yang <fyang@redhat.com>

The existed code will generate an incorrect config when smp
number is small, thus add some check to deal with this situation.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

For SRIOV cases guest mac address will change after reboot.
So need update mac address for them.
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

this patch add test for check groups when guest is started as nobody
[v2] using qemu -runas option to exec the qemu.
Acked-by: NSuqin Huang <shuang@redhat.com>

add support for test check_roms, can use default roms
or some special roms.
Signed-off-by: NShuping Cui <scui@redhat.com>

USB EHCI controller has a 'freq' option, it's used to specify controller
Wakeup frequency.  Effect is that ehci will batch the frame processing
and process 4 frames per wakeup.  Reduces the cpu load, but also reduces
usb performance.  Not guest-visible.

The default value of this option is 1000, which equals the usb frame rate.
Can be reduced to -- for example -- 250.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

If setting machine type is not available, warning message will print
out and default machine type will be used.

Changes from v1:
Add invalid_type parameter.
if we want to test invalid type, need set it to true.
Signed-off-by: NFeng Yang <fyang@redhat.com>

This is a quick fix to make spice work on rhel5 host.
As spice in rhel6 host and rhel5 host is different, we'll
create different functions to add spice in rhel5 hosts.
Signed-off-by: NFeng Yang <fyang@redhat.com>
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

We used to use qemu-ifup to manage the tap which have several
limitations:

1) If we want to specify a bridge, we must create a customized
qemu-ifup file as the default script always match the first bridge.
2) It's hard to add support for macvtap device.

So this patch let kvm subtest control the tap creation and setup then
pass it to qemu-kvm. User could specify the bridge he want to used in
configuration file.

The original autoconfiguration was changed by private bridge setup.

Changes from v1:
* Combine the private bridge config and TAP fd in one patchset,
dropped the "auto" mode

Changes from v2:
* Close TAP fds on VM.destroy() (thanks to Amos Kong for finding
the problem, and for Cleber and Jason for actually fixing it).

Changes from upstream:
* Keep network scripts method and make it configurable
Signed-off-by: NJason Wang <jasowang@redhat.com>
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

Add 'vhost' parameter for vhost=on configure,
and release 'netdev_extra_params' for other extra usage.
Signed-off-by: NGolita Yue <gyue@redhat.com>

Add support of bootindexes for nic dev.

changes from v1:
1) add the bootindex param in add_nic func as last one
Signed-off-by: NShuping Cui <scui@redhat.com>

Signed-off-by: NChen Cao <kcao@redhat.com>

Update add_drive with more parameters using in test. And also update
related part code when generate qemu command line for harddisk, cdrom
and floppy.
Signed-off-by: NFeng Yang <fyang@redhat.com>
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>
Signed-off-by: NYunping Zheng <yunzheng@redhat.com>
Signed-off-by: NShuping Cui <scui@redhat.com>
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

The existed qemu_vm code just verifies if vm.monitors
is not empty, but the property could still return a
None type.

Since in the VM's monitor method there is some code to
validate the vm.monitors member, let's verify the
vm.monitor directly when we accessing it.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

As using gzip to migrate in exec mode requires start the destination
guest after generate the migration file with monitor command line.
Update the code to make migrate.exec in local works again.
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

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

Signed-off-by: Nyunzheng <yunzheng@redhat.com>

Add the guest boot up measure functions to help system_reset
case can make sure reset the guest during the guest boot up.
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

Make pci_addr can be get from cfg and or generate by the
make_create_command.
Signed-off-by: NChen Cao <kcao@redhat.com>

To avoid possible conflict with other tests.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>
Acked-by: NXu Tian <xutian@redhat.com>

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

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

Signed-off-by: NQingtang Zhou <qzhou@redhat.com>
Acked-by: NXu Tian <xutian@redhat.com>

This patch creates a series of functions for subtests in cdrom
testing, and add some parameters to enable/disable these subtests.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

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

Create a separated function to deal with monitor commands,
And add the missing docstring.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

Break the for-loop if we find the matched file name, and slightly
update the code flow to let this function have a single exit.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>
Acked-by: NXu Tian <xutian@redhat.com>

Rename this function to get_empty_cdrom_device, to make its name
match what it does.
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>