Merge pull request !2 from wangshengfang/strace_fix

Merge pull request !3 from wangshengfang/boot_cpu_model_fix

Signed-off-by: NWang Shengfang <wangshengfang@huawei.com>

Signed-off-by: NWang Shengfang <wangshengfang@huawei.com>

Increase sub_test_wait_time to ensure memory device hotplug/uplug when VM
is paused.

Signed-off-by: Wang Ziming <wangziming.huawei.com>

Supported max vcpu less then three times of host cpu number.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

Add more vcpu hot-plug test cases, such as hot-plug 32, 64 or more vcpus.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

Arm machine can run these testcases, when testing large format VM, in case
of out of memory happened, set VM memory as 16384 MiB。Besides, aarch64 vm
vcpu topology only support 'threads=1'.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

Can not offline all vcpus of vm, keep one vcpu alive at least.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

Some machines do not contain irq(15) in file '/proc/irq', need verify
another one if irq(15) not exits.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

autotest module is no longer supported, remove it from test.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

aarch64 only support "device_add" qmp command, suggest using
vm.hotplug_vcpu_device instead of vm.hotplug_vcpu, this modify also need
to change vm configuration before start up.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

aarch64 do not support vcpu hot-unplug, remove it temporary.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

The value of machine_type getted from machines.cfg is incorrect on arm, it should be cut before used.
Signed-off-by: NJie Tang <tangjie18@huawei.com>

The disk device qemu command line has two configurations:
   1). blockdev config, eg:
       -blockdev node-name=file_image1,driver=file,filename=/var/**.qcow2
       -blockdev node-name=drive_image1,driver=qcow2,file=file_image1

   2). disk drive config, eg:
       -drive id=drive_image1,if=none,format=qcow2,file=/var/***.qcow2

If disk drive config, the expect value should be:
 {'device': 'drive_image1', 'qdev': 'image1', 'format': 'qcow2'}

If blockdev config, the expect value should be:
 {'node-name': 'drive_image1', 'qdev': 'image1', 'format': 'qcow2'}
Signed-off-by: NChen Qun <kuhn.chenqun@huawei.com>

Edit threshold adapt to arm64-pci machine type, to make VM memory size as
expected.

Signed-off-by: Wang Ziming <wangziming.huawei.com>

The qemu version is different, some scenes behave differently.
We need to distinguish and judge it, so add the qemu_version.version_compare().
Signed-off-by: NChen Qun <kuhn.chenqun@huawei.com>

Set maxnode from 128 to 32 for openEuler.
Signed-off-by: NChen Qun <kuhn.chenqun@huawei.com>

It will create vm failed, if only add 'vcpu_maxcpus'.
We delete it, and uses 'smp_override' and 'mem_override' params.
Signed-off-by: NChen Qun <kuhn.chenqun@huawei.com>

The check cmd is "qemu-kvm  -M virt -watchdog '?'" for aarch64.
Signed-off-by: NChen Qun <kuhn.chenqun@huawei.com>

add blockdev backup sync top test

induce throttling by opening chardev but not reading data from it

Signed-off-by: NMicky Yun Chan <michan@redhat.com>

savevm_loadvm: save VM and then load it while VM is running

Set smp to vcpu_maxcpus

Add trim test on virtio_scsi

fio_linux: Add fio installation timeout value to make it work on s390x

commit_snapshot_to_raw_backing: add switch for size check 

if image is a block device, the size check will fail the case because
QEMU fails to get correct size info from block device. So add a switch
to enable tester to turn off the snapshot size check.
Signed-off-by: Nlolyu <lolyu@redhat.com>

umount_wipefs: add umount and wipefs steps

qemu_guest_agent: new cast to test fsfreeze hook script

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

add new case: uefi_check_descriptor_meta-files

Verify descriptor meta-files on the host.
Signed-off-by: NXueqiang Wei <xuwei@redhat.com>

Add the timeout value to make the fio installation could finish on
s390x.
Signed-off-by: NNini Gu <ngu@redhat.com>

Add code to be able to transfer the fio installation timeout value from
case level cfg file
Signed-off-by: NNini Gu <ngu@redhat.com>

Fix bug cpu topology: value of sockets * dies * cores * threads
less than smp_cpus.
Signed-off-by: Nqingwangrh <qinwang@redhat.com>

Remove usage of 52lts compatiblity layer

1.Add umount and wipefs steps before formatting in case the hard drive
contains a mounted filesystem.
2.Remove 'format_disk' I/O sub-testing for serial/removable/io_size
options testing.
Signed-off-by: Nybduan <yduan@redhat.com>

block_iscsi_lvm: New case to install on lvm device with iSCSI target