This fault comes from bug:
https://bugzilla.redhat.com/show_bug.cgi?id=912567

The kernel panic message is:
"""
general protection fault: 0000 [#1] SMP
last sysfs file: /sys/power/state
CPU 1
Modules linked in: bridge fuse uinput autofs4 sunrpc 8021q garp stp llc ipt_REJECT nf_conntrack_ipv4 nf_defrag_ipv4 iptable_filter ip_tables ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 xt_state nf_conntrack ip6table_filter ip6_tables ipv6 sg microcode virtio_balloon virtio_console snd_hda_intel snd_hda_codec snd_hwdep snd_seq snd_seq_device snd_pcm snd_timer snd soundcore snd_page_alloc e1000 i2c_piix4 i2c_core ext4 mbcache jbd2 virtio_blk sr_mod cdrom virtio_pci virtio_ring virtio pata_acpi ata_generic ata_piix dm_mirror dm_region_hash dm_log dm_mod [last unloaded: speedstep_lib]

Pid: 2704, comm: bonobo-activati Tainted: G    B      ---------------    2.6.32-358.el6.x86_64 #1 Red Hat KVM
RIP: 0010:[<ffffffff81197619>]  [<ffffffff81197619>] do_sys_poll+0x269/0x520
RSP: 0018:ffff880119d83b28  EFLAGS: 00010202
RAX: 17e05967908ddc73 RBX: ffff880119d83f04 RCX: 953bc6ebfc12a10d
RDX: ffff880119cfcf00 RSI: ffff880119d83f04 RDI: ffff880119cfcf30
RBP: ffff880119d83f38 R08: 953bc6ebfc12a10e R09: 0000000000000001
R10: 0000001100000003 R11: 0000001300000003 R12: ffff880119d83e7c
R13: ffff880119d83e44 R14: ffff880119cfcf00 R15: ffff880119d83b88
FS:  00007f68e5b8d700(0000) GS:ffff880028300000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 00007fdce1db37d7 CR3: 0000000119c3d000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process bonobo-activati (pid: 2704, threadinfo ffff880119d82000, task ffff880119d80aa0)
Stack:
 ffff88011cb4b080 ffff880100000000 ffff880119d80aa0 00ff88011cb4b080
<d> 00007f68d8005450 0000000000000000 ffff880119d83b88 0000000000000000
<d> 0000000000000000 000000008143cfd7 ffff880119d83df8 0000000028216768
Call Trace:
 [<ffffffff81197160>] ? __pollwait+0x0/0xf0
 [<ffffffff81197250>] ? pollwake+0x0/0x60
 [<ffffffff81197250>] ? pollwake+0x0/0x60
 [<ffffffff81197250>] ? pollwake+0x0/0x60
 [<ffffffff81197250>] ? pollwake+0x0/0x60
 [<ffffffff81197250>] ? pollwake+0x0/0x60
 [<ffffffff81197250>] ? pollwake+0x0/0x60
 [<ffffffff81197250>] ? pollwake+0x0/0x60
 [<ffffffff81197250>] ? pollwake+0x0/0x60
 [<ffffffff810ace0b>] ? sys_futex+0x7b/0x170
 [<ffffffff81197b8c>] do_restart_poll+0x3c/0x60
 [<ffffffff8100b341>] ? int_signal+0x12/0x17
 [<ffffffff8108254f>] sys_restart_syscall+0x1f/0x30
 [<ffffffff8100b072>] system_call_fastpath+0x16/0x1b
Code: 00 00 00 0f 1f 80 00 00 00 00 48 89 de e8 60 ac fe ff 48 85 c0 49 89 c6 0f 84 54 01 00 00 48 8b 40 20 48 85 c0 0f 84 37 01 00 00 <48> 8b 40 38 48 85 c0 0f 84 2a 01 00 00 4d 85 ff 74 18 41 0f b7
RIP  [<ffffffff81197619>] do_sys_poll+0x269/0x520
 RSP <ffff880119d83b28>
"""
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>

AHCI disk is in reality ide-drive disk inside AHCI bus.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Fix signal_repeat variant to boot with the correct bootindex.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

Unit test for the response parsing included.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

This way we can write test code that use HMP commands without worrying
if the actualy monitor is a QMP or HMP monitor.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

This way, other code can easily create/register a VM object manually if
necessary.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

This way it can be reused elsewhere. This keeps the on-demand loading of
vm_type-specific modules.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

Instead of importing every single VM type module, just import the
modules that are going to be used on demand, depending on vm_type.

The only exception by now is libvirt_vm, that is also used for the
libvirtd_restart() call at preprocess().
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

The VM class lookup will be moved somewhere else.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

On my test case, I want to get a paused VM and I won't ever need to
unpause it, so I set paused_after_start_vm=yes. The image check code,
however, was unconditionally unpausing the VM, and this is a bug.

This changes the process() code to pause/unpause the VM only if it was
not yet paused.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>

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

spapr_vscsi is used by ppc64.  sd is used on ARM and other embedded
architectures.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The default machine type is still quite x86-specific, but probing qemu or
libvirt for the supported arch is beyond the scope of this series.  I
left a TODO comment in the code.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Also select the machine type in guest-hw.  This is also useful to specify
versioned machine types in a cartesian config file.  Something like this:

    i440fx:
         variants:
         - i440fx-1.0:
              machine_type=pc-1.0
         - i440fx-1.1:
              machine_type=pc-1.1
         - i440fx-1.2:
              machine_type=pc-1.2

lets you run the test multiple times on all versioned types.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

In principle, there is no reason why JeOS images should be Fedora
images.  The config parser picks any JeOS image that satisfies --arch
and --machine-type requirements, and some arches/machine-types may not
be supported by Fedora.  Thus, move existing Fedora-specific values to
the JeOS.17.64 definition.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

We can now tell autotest the architecture being tested, and use it to filter
unwanted tests.  And we can also make the default OS simply "JeOS".  This
prepares for a world where there will be multiple JeOS images, and the arch
filter will pick the right one.

(And in an ideal world, the arch filter will default to whatever is supported
by the chosen QEMU binary).
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Some tests are x86-specific.  Now that all guests call architectures the
same way in the tuples, filter them away.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Consistent naming for the architecture lets test use i386/x86_64/ppc64
in filters.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

"All architectures are equal, but some are more equal than others".
There's no reason for this, avoid it.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

With this change, you can use "./run -g Fedora -tqemu --list-guests"
to list Fedora guests.

Note that this only applies if you specify a "-g" option.  If you
don't, the full list is still passed.  For this reason, the default
guest OS must be specified manually rather than via the command-line
option parser.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Commit b5db3b8 (standalone test: Print 'almost' full guest OS name,
2013-01-09) explains why this could be a problem:

    Win7.64

    is in fact Win7.64.sp0, so if one puts only Win7.64 in
    the command line, we'll execute both Win7.64.sp0 and
    Win7.64.sp1. Printing a part of the full name avoids
    that.

but we can just remove the `@' from sp0 in order to avoid
this trap.  There is no other difference with the current
list of guests.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

This patch is a "Hello, world!" test for virt tests (qemu).

The goal of this script is to provide autotest newbie
an example for some most frequently used interfaces.

Changes from v2:
 * Make test structure more flat as now the framework
   closes unused ssh sessions at the end of the test
 * Use session.cmd, since both this and cmd_status_output
   will raise exceptions so they'd have to be handled
   anyway
 * Use a more customary "Hello, world!" string
 * Really verify whether the guest outputs the right
   string, to show how to fail a test.
 * Introduce checking VM state.
 * Improve the comments on code.

Changes from v1:
 * Fix a typo in commit message.
Signed-off-by: Nyunzheng <yunzheng@redhat.com>
Signed-off-by: NFeng Yang <fyang@redhat.com>
Signed-off-by: NQingtang Zhou <qzhou@redhat.com>
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

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

Sometimes exit_event can be deleted before screendump thread
finishes. Than it hungs in system forever and autotest never
exits.

This problem was spotted after test:
    virtio_console.spread_2.incorrect_options.boot_nr0
where the screendump thread never finished and additional was spawned
at the beggining of the next thread. This caused autotest hang in the and
as only single screendump thread can exit (due of nature of exit fce).

Additionaly set the screendump thread name for future analysis.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Force exit background threads on thest failure.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

When guest script is already running and the test only connects, every
signal destroys the daemon. This is complex problem of python, where
only the main thread accepts the signals. In order to fix this without
much pain I added test_delete_guest_script() at the test beginning thus
fresh instance of guest script is always executed. This solution seems
to work pretty well.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This test closes the used ports while transfering data and after the
intr_time opens them again. It randomly closes ports on guest side
and host side, send threads and recv threads thus in the long run
it should cover all cases.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

When the port is closed, it's completelly removed from the list
of descriptors. When it's opened, the file descriptor might differs
from the original one. Thus it's necessarily to keep the paths
of the used ports in the Switch thread and in case of reconnection
find the new file descriptor.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

ThSendCheck and ThRecvCheck might fail or freeze when they use
closed port.

This is a preparation for interrupted_loopback.disconnect test
where the ports are closed while transfering data.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Require exact PASS statement for some commands as another PASS/FAIL
line might be matched instead.

Also add some workarounds for stucked ports/python by killing the
worker.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Use read_until_any_line_matches as another thread might write
some debug info right after the PASS/FAIL statement. Than
the test fails even thought it should PASS. There are still
some problems with non-thread-safe-print, the most important
are solved with sys.stdout.write(...\n).
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This function is similar to read_until_last_line_matches() function
but it search all lines, not just the last ones. It's useful for
example in virtio_console test where additional output might occur
before it's analyzed with the last_line reader.

The match function is prioritizing the patterns so first it searches
the output for the last pattern, than for the -2 so when there
is PASS and FAIL in the output and patterns = (PASS, FAIL) it matches
the FAIL first, even thought PASS is first in the output.

Separate lines are matched so it's recommended to use '^' at patterns
beginning. ('^PASS', '^FAIL') in order to match the beginning of the
line.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Loopback test might freeze the whole autotest when sender thread
fills the buffer. During cleanup it tries to send additional data
which is blocked due of full buffer. The only solution I found is
to destroy the VM and inform about this bug.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This patch adds interrupted_transfer suspend variant. This variant
initiates loopback over 2 vio ports and suspends the VM to mem.
After a while it resumes the VM with 'system_wakeup' and verifies
that the data are still transfered in booth directions.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Set more than allowed number of ports per virtio-serial-pci.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

In order to be able to run all virtio smoke tests easily I added
virtio_console_smoke variant with all related subtest inside.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>