balloon service should be in running status after send INTERROGATE
signal via windows cmd.
Signed-off-by: NLi Jin <lijin@redhat.com>

block_hotplug:add code to unplug virtio-scsi-pci.
Signed-off-by: NAihua Liang <aliang@redhat.com>

This test did not pass the "migration_cmd*" to the `vm.migrate`,
therefor the `exec.default_exec` and `exec.gzip_exec` were the same.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

1. When windows guest mem is small(e.g.1024M), calculated mem
min_size will be larger than ori_mem if ratio=0.5, which will
lead balloon cases fail.
2. When do balloon operation in linux guest, we should consider
crashkernel value as well,leaving enough memory for guest to
avoid guest OOM.

So modify the formula and leave a balloon buffer(300M in default)
for guest min_sz.
Signed-off-by: NLi Jin <lijin@redhat.com>

original sleep time(90s) is too small, balloon_check get guest used
memory at the very early stage. The value is not correct as it will
increase after all guest application up.
Signed-off-by: NLi Jin <lijin@redhat.com>

Only virtio driver need to be checked in windows, no need
to do it with ide.
Signed-off-by: NAihua Liang <aliang@redhat.com>

Fio has multiple device function, it allows fio running
on multiple disks at the same time. Here use this way to catch
run fio on system and data disks at the same time.
Signed-off-by: NPeixiu Hou <phou@redhat.com>

Add the missing LF at the end of the cfg file.
Signed-off-by: NXu Han <xuhan@redhat.com>

Make test cases to run with the certain version of qemu
Signed-off-by: NPing Li <pingl@redhat.com>

vm.serial_console is created by default, can be called directly.
Signed-off-by: NSitong Liu <siliu@redhat.com>

The format of pci addr was changed to hex numbers with leading "0x" a
while ago, but this test was not adjusted resulting in reserved address
having unusable addresses. Let's adjust the format.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The pci_devices test needs to insert high-priority devices into
low-priority ones, which is not possible due to automatic sorting. Let's
disable the autosorting.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The timeout 10s was chosed to not slow-down the test execution much, but
usually less than 1s is enough. These days the boot takes < 20s,
sometimes even 5s so let's decrease the timeout to suitable but still
safe 5s.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The "pci_bridge1" became widely used and it clashes with the default
name used by this test. Let's use "test_pci_%s" to emphasize it's
test-defined pci controller.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The make_create_command used to return "devices", but now it returns
"tuple(make_create_devices, spice_options)". Let's allow both by
checking the type.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

add install/uninstall and upgrade/downgrade driver cases for pvpanic
Signed-off-by: NYu Wang <wyu@redhat.com>

Add driver install case with pvpanic device
Signed-off-by: NYu Wang <wyu@redhat.com>

qemu_tree.check_guests_proc_scsi() only compares scsi infos
for qtree/guest, so when there's an ide/virtio_blk device,
call this function is useless.

And it can't distigush a scsi device and a usb device, so
this function call for scsi/usb is also meaningless.
Signed-off-by: NAihua Liang <aliang@redhat.com>

new cases related pvpanic:
1. check pvpanic event after trigger crash by nmi
2. do nmi crash after system_reset/reboot/shutdown/stop_cont
3. do nmi crash after migration
Signed-off-by: NYu Wang <wyu@redhat.com>

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

Signed-off-by: NXiyue Wang <xiywang@redhat.com>

In boot_from_device.check_boot_result(), although console output
boot_fail msg, the test result will be set to pass. The logic is
not reasonable, correct it.
Signed-off-by: NAihua Liang <aliang@redhat.com>

"cmd = WINUTILS:\hddsn.exe C:" is not match the lable name
in def set_winutils_letter, change it to WIN_UTILS.
Signed-off-by: NPeixiu Hou <phou@redhat.com>

The mmio-based arm requires virtio-serial-device and s390x machine
requires virtio-serial-ccw device.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

test errors out if tcpdump is not installed in guest, handle it by
checking and installing it.

remove few autotest dependencies
Signed-off-by: NBalamuruhan S <bala24@linux.vnet.ibm.com>

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

get_memory_boundary function will calculate a min_size for balloon operation.
Default ratio value(0.1) is too small which lead to small min_size.Actuall
windows minimal mem increases if guest stress increase, sometimes it makes
balloon operation fail. Set ratio = 0.5 in the cfg file to avoid this issue.
Signed-off-by: NLi Jin <lijin@redhat.com>

when sub_test is a cmd instead of a case in avocado, use
session.cmd() function to run it, rather than
utils_test.run_virt_sub_test()
Signed-off-by: Nlijin <lijin@redhat.com>

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

Current 1800s is not long enough to run 1000 times balloon on some hosts. Remove the config to use the default value in base.cfg.
Signed-off-by: NYumei Huang <yuhuang@redhat.com>

Add a scsi-hd device for system disk on ppc and modify the strategy of checking interrupts.
Signed-off-by: NYongxue Hong <yhong@redhat.com>

Signed-off-by: NHaotong Chen <hachen@redhat.com>

The "drive_mirror_stress" uses "error_context" without setting
"context_aware" which leads to error. Let's set it.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This improves the "cpuflags-test" exception message and adds second
check for the "stressblock" file being created by background "dd"
command (which used not to be created due to test error).
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This adds checks for at least one verified flag being available and no
flags failing the check before the migration.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The `if=/dev/[svh]da` can't be evaled by bash due to `if=` prefix. Let's
use `$(echo /dev/[svh]da)` instead.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The parent class's cleanup takes care of the "self.vm" and unlike this
method works on earlier failures as well.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This swap is not necessary and makes it less clear whether we act on vm
or clone. Let's be more explicit and rely on "cleanup" to shutdown both.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Let's clean both "vm" and "clone" if they exists to avoid complex
modifications required to switch from "vm" to "clone".
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The "cpu_flags" dependency was moved which changed in
074fd49c and
567d54dc but not all paths were
addressed. Let's take care of the rest of the paths.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>