In some enterprises, the default bootindex is -1 for system disk,
which means cancel the boot index.
If hotplug a data disk then reboot guest, qemu may set the data
disk with higher bootindex, which depends on different enterprises'
policies.
As above, need set bootindex=0 for system disk.
Signed-off-by: NCong Li <coli@redhat.com>

Read data from /dev/random 10 hours
Signed-off-by: NSuqin Huang <shuang@redhat.com>

Add qemu-iotests cases for qcow2/raw/vmdk/qed/vhdx image format
Signed-off-by: NPing Li <pingl@redhat.com>

1. repeat load/unload 10 times
2. unload during random read
Signed-off-by: NSuqin Huang <shuang@redhat.com>

According to new policy,all RHEL host will use the latest virtio-win drivers.
And latest virtio-win serial driver use "VirtIO Serial Driver".
Signed-off-by: NYanan Fu <yfu@redhat.com>

Signed-off-by: NYumei Huang <yuhuang@redhat.com>

Signed-off-by: NYumei Huang <yuhuang@redhat.com>

Signed-off-by: NXu Tian <xutian@redhat.com>

RHEL.6 host does not support "ich9-usb-uhci6".

tests.netperf_stress:
1.add params of "prompt,linesep, status_test_command"
2.pass these params to "NetperfServer/NetperfClient"

tests.netperf_udp:
1.add params of "prompt,linesep, status_test_command"
2.pass these params to "NetperfServer/NetperfClient"

tests.flow_caches_stress_test:
1.add params of "prompt,linesep, status_test_command"
2.pass these params to "NetperfServer/NetperfClient"

tests.migration_with_netperf:
1.add params of "prompt,linesep, status_test_command"
2.pass these params to "NetperfServer/NetperfClient"

tests.multi_nics_stress:
1.add params of "prompt,linesep, status_test_command"
2.pass these params to "NetperfServer/NetperfClient"

tests.ovs_qos:
1.add params of "prompt,linesep, status_test_command"
2.pass these params to "NetperfServer/NetperfClient"
3.correct info index for status_test_command

Signed-off-by:Aihua Liang<aliang@redhat.com>

Add case balloon_boot_in_pause to balloon guest memory when guest started in paused status

guest didn't enforce quit while flag lacked in RHEL.6 host, and this problem
just fix on RHEL.7 host, so disable this case on RHEL.6 host.
Signed-off-by: NXueqiang Wei <xuwei@redhat.com>

1. separate live snapshot negative tests from live_snapshot.cfg to a new configuration.
2. rename live_snapshot_negative_test.py to live_snapshot_negative.py for a uniform format.
   And also rename the class name in it.
3. update error match string for RHEL.6 host.
Signed-off-by: NCong Li <coli@redhat.com>

block job pause and resume for storage migration are not supported on RHEL.6 host.
Signed-off-by: NCong Li <coli@redhat.com>

xhci is not supported on RHEL.6 host.
Signed-off-by: NCong Li <coli@redhat.com>

qemu-img should not create image when cluster_size=0.
There are some problems of cluster_size=0 on RHEL.6 host and which will not be fixed,
so remove this scenario on RHEL.6 host.
Signed-off-by: NCong Li <coli@redhat.com>

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1324342

1. case logic is wrong, there are 2 types in the configuration, 1 is balloon_check, the other is kernel_install, actually kernel_install is the case which will be executed.
2. no need to match qmp output pattern now.
3. balloon_check_qmp steps can be covered in balloon_check with qmp, and no scenario about kernel_install and balloon_check.
Signed-off-by: NCong Li <coli@redhat.com>

1. block_hotplug.cfg: add separate scenario for repeated plug/unplug, no need repeat 300 times for each scenario when one pci.
2. block_hotplug.py: get and close a separate session when using it to fix file descriptor out of range problem.
3. stop_continue.py: close session when test finished.
Signed-off-by: NCong Li <coli@redhat.com>

When drive mirroring with qcow2 image on iscsi, it will cause mismatch issue, which is caused by iscsi instead of qemu,
so only test raw image on iscsi, qcow2 matrix will be covered with local and nfs storage.
Signed-off-by: NQianqian Zhu <qizhu@redhat.com>

Signed-off-by: NYumei Huang <yuhuang@redhat.com>

need to read from /dev/random instead of /dev/hwrng

/dev/random is the place from where you should always read
/dev/hwrng will be available to read if you have a hardware rng
device
Signed-off-by: NSuqin Huang <shuang@redhat.com>

1. usb-ehci controller can not connect with low-speed usb device when machine type is rhel6.* .
2. In base.cfg, default add ich9-usb-ehci1 controller, so redefine it in this scenario on RHEL.6 host.
Signed-off-by: NCong Li <coli@redhat.com>

Signed-off-by: NXu Han <xuhan@redhat.com>

Signed-off-by: NXu Han <xuhan@redhat.com>

block job test will eat a lots of caches in testing,
free caches  after test to avoid VM hang memory.
Signed-off-by: NXu Tian <xutian@redhat.com>

BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1400385Signed-off-by: NJing Zhao <jinzhao@redhat.com>

Signed-off-by: NCong Li <coli@redhat.com>

Signed-off-by: NPing Li <pingl@redhat.com>

Signed-off-by: NQianqian Zhu <qizhu@redhat.com>

Signed-off-by: NYanan Fu <yfu@redhat.com>

Signed-off-by: NQianqian Zhu <qizhu@redhat.com>

Signed-off-by: NQianqian Zhu <qizhu@redhat.com>

1. fix typo in single_driver_install.cfg.
2. add q35 configurations in boot_order_check.cfg.
Signed-off-by: NCong Li <coli@redhat.com>

sleep some seconds before check ballooned memory
Signed-off-by: NAihua Liang <aliang@redhat.com>

Signed-off-by: NXu Tian <xutian@redhat.com>

Signed-off-by: NWei Liao <weliao@redhat.com>

Fix a typo, verfiy -> verify; add a preposition 'with'.

Signed-off-by: NCong Li <coli@redhat.com>