A frequent request we've got from users is that the name
'kvm' makes assumptions that are not true for a lot of
the use cases of the test suite. Then let's rename
the subtest 'kvm' to 'qemu', making it more accurate.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Clean not only the environment file, but also the
address_pool and the temporary aexpect_spawn directory
on every test run.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

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

Signed-off-by: NMarian Krcmarik <mkrcmari@redhat.com>

Signed-off-by: NMarian Krcmarik <mkrcmari@redhat.com>

Signed-off-by: NMarian Krcmarik <mkrcmari@redhat.com>

Instead of guessing them from xrandr output, which doesn't
display values in available xrandr modes.
Signed-off-by: NMarian Krcmarik <mkrcmari@redhat.com>

Minor fixes to suppress console output with some escaped
characters, ingore some exceptions often hit but being
false positive when possible.
Signed-off-by: NMarian Krcmarik <mkrcmari@redhat.com>

- fix condition of wxPython installation
- send useless cmd output to /dev/null
- use launch_startx from utils_spice
Signed-off-by: Nroot <root@dhcp131-204.brq.redhat.com>

Signed-off-by: NMarian Krcmarik <mkrcmari@redhat.com>

Adding function widely used in rv_* tests and
ingoring some exceptions which are most likely caused by
color console output.
Signed-off-by: NMarian Krcmarik <mkrcmari@redhat.com>

Assigning "None" to spice_port parameter does not lead to
generating a free spice port when specific value has been
assigned in previous tests of one autotest run.
Signed-off-by: NMarian Krcmarik <mkrcmari@redhat.com>

Normalizing the variant names to be all lowercase
and use underscores to be consistent throughout the file.
Added the logging tests for spice (spice_vdagent and qxl).
Signed-off-by: NVimal Patel <vipatel@redhat.com>

rv_logging will test the logging of the qxl driver and spice_vdagent daemon on the guest machine.
Signed-off-by: NVimal Patel <vipatel@redhat.com>

1) vm.create() have to force new params.
2) use data_dir instead of vm.root_dir for image absolute path
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Add a test which plugs and unplugs (virtual) usb devices.
Configured to run with the same set of devices we do
boot+reboot tests with.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Add test for usb-test (aka pass through).  Requires a physical usb device,
which must be specified in the configuration.  When unconfigured the test
raises TestNAError so the runner flags the test as skipped.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Create a new "usb" test and move over all usb subtests step by step.
Reduce duplication and allow to run usb tests with a simple --tests=usb
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

If an exception happened when assembling the qemu cmd
(for example, unsupported cmdline flags for that
particular qemu version), clean up any tap fd open
for each nic.

This is the more proper solution for the issue first
mentioned here:

https://github.com/autotest/autotest/issues/520

That Gerd hit while running the usb related tests
in the code using RHEL 6. This patch finally makes
the usb tests properly pass when using bridge:

[root@virtblade02 virt-test]# ./run -t kvm --nettype bridge --tests "boot_with_usb reboot_with_usb"
SETUP: PASS (1.55 s)
DATA DIR: /var/lib/virt_test
DEBUG LOG: /root/virt-test/logs/run-2012-12-24-00.27.31/debug.log
TESTS: 20
(1/20) boot_with_usb.ehci: PASS (22.82 s)
(2/20) boot_with_usb.keyboard.uhci: PASS (27.84 s)
(3/20) boot_with_usb.keyboard.xhci: SKIP
(4/20) boot_with_usb.mouse.uhci: PASS (26.95 s)
(5/20) boot_with_usb.mouse.xhci: SKIP
(6/20) boot_with_usb.usb_audio: SKIP
(7/20) boot_with_usb.hub: PASS (27.07 s)
(8/20) boot_with_usb.storage.uhci: PASS (27.76 s)
(9/20) boot_with_usb.storage.ehci: PASS (27.86 s)
(10/20) boot_with_usb.storage.xhci: SKIP
(11/20) reboot_with_usb.ehci: PASS (62.71 s)
(12/20) reboot_with_usb.keyboard.uhci: PASS (64.18 s)
(13/20) reboot_with_usb.keyboard.xhci: SKIP
(14/20) reboot_with_usb.mouse.uhci: PASS (63.01 s)
(15/20) reboot_with_usb.mouse.xhci: SKIP
(16/20) reboot_with_usb.usb_audio: SKIP
(17/20) reboot_with_usb.hub: PASS (63.05 s)
(18/20) reboot_with_usb.storage.uhci: PASS (64.08 s)
(19/20) reboot_with_usb.storage.ehci: PASS (64.16 s)
(20/20) reboot_with_usb.storage.xhci: SKIP

CC: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Make it possible to specify the main vm's memory through
the parameter --mem (shorthand -m).
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

If you need to use bridge and tap devices for some
specific reason, now it is possible to specify those
options through the parser command line.

If you do choose --nettype = bridge however, you must
run the execution script as root, and we enforce that
restriction.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Show on debug messages where we looked for test modules,
what is the search criteria, the modules we found.

Also, put a bit of spacing and some messages that help
to make sense out of the debug log.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Print some useful info in the debug log:

* Timestamp of test job start
* Options passed on the command line
* If using the kvm test, print the qemu binaries used
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

So on every test the pristine JeOS image is copied over
on every new test. We could've implemented something
with internal disk snapshots, but then we'd restrict
ourselves only to disk images that support snapshots.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Now when setting restore_image = yes, the postprocessing
code will *always* restore the image from backup. With
this we can implement an option in the test runner to
always restore the image from backup between tests.

Useful for folks that want to be extra careful, and
ensure the guest image is always pristine and clean
on every new test.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Add a general flag to restart the vm between tests, which
does nothing but set restart_vm = yes for all cartesian
dicts.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Make it possible to choose between tcg and kvm when running
the kvm specific tests. This is part of a series of changes
that will end up in renaming what we know today as 'kvm'
test to 'qemu'.

I think there's a bunch of assumptions that we make throughout
the tests with regards to qemu being accelerated with kvm,
that we'll discover when running the tests with tcg. Stay
tuned.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

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

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

This was one tweak on the autotest control files not
present in the test runner - There are occasions where
localized output of certain commands might cause failures,
so let's ensure the language is en_US.UTF-8.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

So it is easier to visualize each one of these options.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Organize test runner code better under an App class.
This way it is easier to visualize in high level terms
what the test runner does.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Create a specialized class for parsing the arguments,
make the code better organize, and start categorizing
options (such as general options and options specific
to a given virt subtest).
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

'qemu-kvm -device ?' print output to stderr. So have to udpate the
code to make it work.
Signed-off-by: NFeng Yang <fyang@redhat.com>

Used wrong path and thus didn't match image file names.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Currently, one of kvm cgroup test, cpuset_cpus, will get and error
if the number of vcpus not equal to the hardware cpus.
for example, when vm_cpus is 2, and the host cpu number
(no_cpus) is 4, while i >= 3, vcpus[i - 1] will make no sense.

To avoid this, add i <= vm_cpus to correct it.
Signed-off-by: NMike Qiu <qiudayu@linux.vnet.ibm.com>

Enable demo mode of netperf, process mixed demo results to drop
unstable factors, only compute the effective demo results when
all the threads work. This would make performance result more
stable.

We start many clients and watch the demo output, if all the threads
really work, then start to record the interrupt/packet number, and
count the cpu usage. After a fixed test time, record the final
interrupt/packet number.
Signed-off-by: NAmos Kong <akong@redhat.com>

Currently we use '-b' parameter of netperf to test parallel traffic,
client would send many requests at start of _RR tests, but it's not
real parallelism, we want to use demo mode to improve the
stabilization. This patch removes '-b' parameter and uses multiple
threads for TCP_RR testing.
Signed-off-by: NAmos Kong <akong@redhat.com>

Rename 'client' to 'clients'
Signed-off-by: NAmos Kong <akong@redhat.com>