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

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

Signed-off-by: NLi Yang <liyang.fnst@cn.fujitsu.com>

Introduce an extra parameter to the function, an
output file. If it is passed, write the output of
the command to a file. This is necessary for the
virsh net-list test.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Signed-off-by: NLi Yang <liyang.fnst@cn.fujitsu.com>

Signed-off-by: NLi Yang <liyang.fnst@cn.fujitsu.com>

Test command: virsh change-media.

The command changes the media used by CD or floppy drives.

Test steps:
1. Prepare test environment.
2. Perform virsh change-media operation.
3. Recover test environment.
4. Confirm the test result.

Changes from v1:

 * Replaced needless use of shell commands
   when there is a valid python API that can
   be used instead
 * Fixed check of files inside a cdrom device
   inside the guest
 * Normalized references to CDrom to cdrom.
 * Marked tests requiring libvirt restart to
   require_root
 * Squashed test and config on a same commit
Signed-off-by: Nwhuang <whuang@redhat.com>

Signed-off-by: Nwhuang <whuang@redhat.com>

Signed-off-by: NYanbing Du <ydu@redhat.com>

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

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

As mentioned, it'll be better and easier to have as much
as possible of the arch dependent code in this module.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

In order to concentrate (as much as possible) processor
architecture dependent code, create an arch module on
virt tests. As the first function in there, refactor
the code that gets the list of kvm modules on different
architectures.

The idea is to move other arch dependent code to that
module, so we have a cleaner and less repetitive way
of solving multi-arch support.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Currently, check_modules in qemu/get_started.py and
virttest/standalone_test.py just suitable for x86_64, for other
arch, especially ppc64, will not have module like kvm-%s, and this
causes errors.

To fix this problem, use platform to decide the modules need to
be checked.
Signed-off-by: NMike Qiu <qiudayu@linux.vnet.ibm.com>

Currently, the autotest haven't support the guest os fedora17
and 18 for ppc64.

Add two config file to support this to os.
Signed-off-by: NMike Qiu <qiudayu@linux.vnet.ibm.com>

Currently, the regular expression in extract_qemu_cpu_models just
works with X86 qemu models.

Add check if X86 pattern return none, will use PowerPC pattern to
get qemu support cpu model.

If it is not possible to recognize the patterns in
the output of 'qemu -cpu ?', throw an error explaining
that the proper support must be added.
Signed-off-by: NMike Qiu <qiudayu@linux.vnet.ibm.com>

Currently, virt-tests support amd and intel's X86 cpu type,
IBM Power cpus will be detected as 'unknown'.

To solve the problem, use platform.machine, so when the
arch is ppc64, we can return appropriate power CPU models.
Signed-off-by: NMike Qiu <qiudayu@linux.vnet.ibm.com>

Use funcatexit to handle the debugfs clean up in transparent hugepage test
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

With this module, tests can define multiple clean up functions which will
be called after tests termination inside the tests scripts. These will help
tests which need complex host enviroment setup and clean up.

The clean up function can be defined in tests and will be register to a
global list named with exithandlers of the test. You can register the
clean up functions right after the set up operation finished. And they
will be executed no matter the test is failed or not after the test finished.
And the execution order will be the backward of the register order.

Changes from v1:
 * Fixed typo, removed unused module and trailing whitespaces.
Signed-off-by: NYiqiao Pu <ypu@redhat.com>

This patch benefits from the better error message when key is missing
in config params. Instead of custom handling of missing param it
uses params[] directly.

I modified those params, where I was 100% certain that they are
required and doesn't have default value:
 * val = params.get(..) followed by if not val: raise...
 * int(params.get(...))
 * params.get(...).split()
 * commands which require string (eg. session.cmd(params.get(...)))

I didn't test all the tests before and after in order to verify
possible typos but I used pylint before and after, watching for
new errors (code rate before and after).
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

They'll come up handy, and will help to verify whether
the Params class is behaving like we intended.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

This error message covers most cases of the missing config key thus
it allow us to use params[] for required configs instead of checking
and raising custom exceptions.

Changes from v1:
 * Raise a custom error, ParamNotFound, that inherits
   from error.TestNAError, that will make the tests
   that lack a mandatory param to be skipped.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

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

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

Several virsh tests share a need in executing a virsh command on
a remote host, against the virt test client machine.

The VirshSession class was enhanced to accept remote-access
parameters.  When present, an ssh command is executed prior
to entering the virsh shell (on the remote system).  When
these parameters aren't provided, this class operates as before.

A new VirshConnectBack class was added as a subclass of
VirshPersistent.  It's only override is to validate the
remote-access parameters are provided, then create the
VirshSession using them.
Signed-off-by: NChris Evich <cevich@redhat.com>

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

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

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

Useful to see details of the config update process.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

When bootstrap is run once, all files are updated
properly. However, as it runs on subsequent times,
modifications in the base cfg file will be ignored,
since the files were already populated. So, add the
additional criteria that the base cfg files can only
be overriden if the test specific file is git tracked.

This way, */update_config.py will always produce the
desired effect (actually replacing the local test
cfg files with the new, modified files).
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Signed-off-by: NSatheesh Rajendran <sathnaga@linux.vnet.ibm.com>

Signed-off-by: NSatheesh Rajendran <sathnaga@linux.vnet.ibm.com>

Signed-off-by: NSatheesh Rajendran <sathnaga@linux.vnet.ibm.com>

Despite what I said in ff75623b17a4c690ed7b098d74149b8a642066b1,
we weren't actually shared qemu/tests/cfg/unittest.cfg. Fix it.
Signed-off-by: NCole Robinson <crobinso@redhat.com>

When self.make_create_command() raises an exception in VM.create()
it's catched and reraised. The problem is that the information about
actual failed line is lost. This patch logs the original traceback
and raise VMStartError instead.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This is a copy of function from autotest.client.shared.base_utils.
Keep it here for few weeks in order to allow users with old autotest
run the current virttest.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

int(params.get(...), $int) should be probably default value of ...
=> int(params.get(..., $int))
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Adds file which contain patterns which should be excluded
from tar package distributed to clients from the server.
Signed-off-by: NJiří Župka <jzupka@redhat.com>

iSCSI test supports only tgtd implementation. But there is prepared
interface for adding other type of iSCSI server.
Signed-off-by: NJiří Župka <jzupka@redhat.com>

param:
  images_base_dir = "new path"
Signed-off-by: NJiří Župka <jzupka@redhat.com>