[KVM-AUTOTEST] Patch queue for kvm autotest

[KVM-AUTOTEST] Patch queue for kvm autotest

[Lucas Meneghel Rodrigues]

Since we've started work on upstream merge, there's a wealth of patches
that were not integrated to kvm autotest that now needs to be re-worked
and eventually integrated to the upstream kvm test.

I've made a list, and I am copying the patch authors, so we make a task
force to re-work and integrate the outstanding patches, helping to make
our transition to upstream development smoother. I would like to help on
whatever is needed to get the test code in good shape so we can move on
to grow our functional test base.

We have several patches already reviewed and fairly straightforward to
rebase/apply. I am willing to rebase them and already send to the
mailing list. If I hear no further comments on them, they will be
applied. I mark them as 'Low hanging fruit' on the list below. Other
patches need a bit more work on them, so I marked them as 'Need rework'.
When I felt necessary, I added some comments. Allways when possible,
I've put a kerneltrap mailing list archive URL to the patch summary.

We have a git tree that mirrors autotest upstream, so feel free to clone
it to rework your patches, I believe it's the fastest way to start doing
it:

http://github.com/ehabkost/autotest/tree/master

I sincerely apologize for any inconveniences the merge upstream has
caused, and let's work hard to get rid of this backlog. Hopefully soon
things will be much smoother for everyone.

Patch queue:

--------------------------------------------------------------------------------
David Huff (dhuff@redhat.com)

1) Move kvm functional tests to a 'test' directory

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/26/5812453

For the sake of organization and extensibility, this is one of the
things I would like to see done. Something that does not require us to
explicitely register tests on the kvm test class would be handy, and
having all tests under a 'tests' directory seems a Good Thing (TM) to
pursue.
Needs rework

2) Fix to allow for "=" in the "value" of a config parameter

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/21/5790533/thread

Low hanging fruit

--------------------------------------------------------------------------------
Yogi (anantyog@linux.vnet.ibm.com)

1) Support for remote migration

http://kerneltrap.org/mailarchive/linux-kvm/2009/4/30/5607344/thread

This is one of the things we should pursue. Remote migration needs to be
stressed out then I ask Yogi to rework his patch, and would like to help
to get this in good shape for inclusion.
Needs rework

--------------------------------------------------------------------------------
Bear Yang (byang@redhat.com)

1) Time drift test

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/13/5778283/thread

After some discussion with Yaniv Yakul (ykaul@redhat.com) and Bear, it
was uncertain to me what to do next. I am all for including it, and
indeed we can drop the ntp synchronization between hosts.
Needs rework

--------------------------------------------------------------------------------
Lucas Rodrigues (lmr@redhat.com)

1) Install refactor

http://kerneltrap.org/mailarchive/linux-kvm/2009/6/3/5868753/thread

Already re-worked against upstream. Needs to be reviewed.

--------------------------------------------------------------------------------
Michael Goldish (mgoldish@redhat.com)

1) RHEL-4.7 step files: fix the initial boot barriers

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797903/thread

Low hanging fruit

2) WinXP step files: add an optional barrier to deal with a closed start
menu

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797913/thread

Low hanging fruit

3) RHEL-5.3 step files: fix initial boot barriers and an inconsistent
dialog

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798843/thread

Low hanging fruit

4) stepeditor.py: get rid of some annoying keyboard shortcuts

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797953/thread

Low hanging fruit

5) kvm_vm.py: choose a monitor filename in the constructor instead of
VM.create()

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797943/thread

Low hanging fruit

6) kvm_vm.py: add new function VM.clone() which returns a clone of the
VM

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798283/thread

This patch is fine, even with the larger code proposed initialy by
Michael.
Low hanging fruit

7) kvm_vm.py: make sure the bulk of VM.create() is not executed in
parallel

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798713/thread

Low hanging fruit

8) kvm_runtest_2.py: use pickle instead of shelve when loading/saving
env

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798483

Low hanging fruit

9) kvm_runtest_2.py: use environment filename specified by the 'env'
parameter

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798473/thread

Low hanging fruit

10) kvm_vm.py: add new VM parameter 'x11_display' that controls $DISPLAY

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798493

Low hanging fruit

11) VM.create(): always destroy() the VM before attempting to start it

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797933

Low hanging fruit

12) kvm_utils.py: remote_login(): improve regular expression matching

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798273

Low hanging fruit

--------------------------------------------------------------------------------
Supriya Kannery (supriyak@in.ibm.com)

1) Iterate over reboot

http://kerneltrap.org/mailarchive/linux-kvm/2009/4/20/5529074/thread

I revised the thread, saw that Uri made some modifications to the patch
made by Supriya, but couldn't find a reference to this on the git log,
so I am not sure if we already have this integrated.

? Needs rework ?

--------------------------------------------------------------------------------
Yolkfull Chow (yzhou@redhat.com)

1) fix a little problem for kvm_config.py

Bugfix, easy to apply. Shouldn't require a lot of rework.

Low hanging fruit

--------------------------------------------------------------------------------
Mike Burns (mburns@redhat.com)

1) Make code to remove kvm modules more robust

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/20/5787593/thread

I have re-worked this patch from Mike and proposed upstream

http://patchwork.test.kernel.org/patch/650/

Since it's taking a bit longer than expected to review by upstream (will
poke Martin), I have put it under kvm_utils, on my install refactor
patch just for precaution. Once it gets into utils, I am going to drop
the code on kvm_utils. Bottom line, I have assimilated this patch on
other patches of mine (with Mike's permission of course).

Assimilated into other patches

--------------------------------------------------------------------------------
Alexey Eromenko (aeromenk@redhat.com)

1) New test module: iperf

http://kerneltrap.org/mailarchive/linux-kvm/2009/5/31/5840973/thread

Needs rework

--------------------------------------------------------------------------------
Jason Wang (jasowang@redhat.com)

1) TAP network support in kvm-autotest

I didn't find this one on the kernel trap archives

Needs rework


-- 
Lucas Meneghel Rodrigues
Software Engineer (QE)
Red Hat - Emerging Technologies

Re: [KVM-AUTOTEST] Patch queue for kvm autotest

[Yaniv Kaul]

On 6/5/2009 8:46 PM, Lucas Meneghel Rodrigues wrote:
> Since we've started work on upstream merge, there's a wealth of patches
> that were not integrated to kvm autotest that now needs to be re-worked
> and eventually integrated to the upstream kvm test.
>
> I've made a list, and I am copying the patch authors, so we make a task
> force to re-work and integrate the outstanding patches, helping to make
> our transition to upstream development smoother. I would like to help on
> whatever is needed to get the test code in good shape so we can move on
> to grow our functional test base.
>
> We have several patches already reviewed and fairly straightforward to
> rebase/apply. I am willing to rebase them and already send to the
> mailing list. If I hear no further comments on them, they will be
> applied. I mark them as 'Low hanging fruit' on the list below. Other
> patches need a bit more work on them, so I marked them as 'Need rework'.
> When I felt necessary, I added some comments. Allways when possible,
> I've put a kerneltrap mailing list archive URL to the patch summary.
>
> We have a git tree that mirrors autotest upstream, so feel free to clone
> it to rework your patches, I believe it's the fastest way to start doing
> it:
>
> http://github.com/ehabkost/autotest/tree/master
>
> I sincerely apologize for any inconveniences the merge upstream has
> caused, and let's work hard to get rid of this backlog. Hopefully soon
> things will be much smoother for everyone.
>
> Patch queue:
>
> --------------------------------------------------------------------------------
> David Huff (dhuff@redhat.com)
>
> 1) Move kvm functional tests to a 'test' directory
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/26/5812453
>
> For the sake of organization and extensibility, this is one of the
> things I would like to see done. Something that does not require us to
> explicitely register tests on the kvm test class would be handy, and
> having all tests under a 'tests' directory seems a Good Thing (TM) to
> pursue.
> Needs rework
>
> 2) Fix to allow for "=" in the "value" of a config parameter
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/21/5790533/thread
>
> Low hanging fruit
>
> --------------------------------------------------------------------------------
> Yogi (anantyog@linux.vnet.ibm.com)
>
> 1) Support for remote migration
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/4/30/5607344/thread
>
> This is one of the things we should pursue. Remote migration needs to be
> stressed out then I ask Yogi to rework his patch, and would like to help
> to get this in good shape for inclusion.
> Needs rework
>
> --------------------------------------------------------------------------------
> Bear Yang (byang@redhat.com)
>
> 1) Time drift test
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/13/5778283/thread
>
> After some discussion with Yaniv Yakul (ykaul@redhat.com) and Bear, it
> was uncertain to me what to do next. I am all for including it, and
> indeed we can drop the ntp synchronization between hosts.
> Needs rework
>    
Michael Goldish is implementing it. He's been on vacation, but I expect 
it to be ready pretty soon.
Y.

> --------------------------------------------------------------------------------
> Lucas Rodrigues (lmr@redhat.com)
>
> 1) Install refactor
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/6/3/5868753/thread
>
> Already re-worked against upstream. Needs to be reviewed.
>
> --------------------------------------------------------------------------------
> Michael Goldish (mgoldish@redhat.com)
>
> 1) RHEL-4.7 step files: fix the initial boot barriers
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797903/thread
>
> Low hanging fruit
>
> 2) WinXP step files: add an optional barrier to deal with a closed start
> menu
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797913/thread
>
> Low hanging fruit
>
> 3) RHEL-5.3 step files: fix initial boot barriers and an inconsistent
> dialog
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798843/thread
>
> Low hanging fruit
>
> 4) stepeditor.py: get rid of some annoying keyboard shortcuts
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797953/thread
>
> Low hanging fruit
>
> 5) kvm_vm.py: choose a monitor filename in the constructor instead of
> VM.create()
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797943/thread
>
> Low hanging fruit
>
> 6) kvm_vm.py: add new function VM.clone() which returns a clone of the
> VM
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798283/thread
>
> This patch is fine, even with the larger code proposed initialy by
> Michael.
> Low hanging fruit
>
> 7) kvm_vm.py: make sure the bulk of VM.create() is not executed in
> parallel
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798713/thread
>
> Low hanging fruit
>
> 8) kvm_runtest_2.py: use pickle instead of shelve when loading/saving
> env
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798483
>
> Low hanging fruit
>
> 9) kvm_runtest_2.py: use environment filename specified by the 'env'
> parameter
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798473/thread
>
> Low hanging fruit
>
> 10) kvm_vm.py: add new VM parameter 'x11_display' that controls $DISPLAY
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798493
>
> Low hanging fruit
>
> 11) VM.create(): always destroy() the VM before attempting to start it
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5797933
>
> Low hanging fruit
>
> 12) kvm_utils.py: remote_login(): improve regular expression matching
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/24/5798273
>
> Low hanging fruit
>
> --------------------------------------------------------------------------------
> Supriya Kannery (supriyak@in.ibm.com)
>
> 1) Iterate over reboot
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/4/20/5529074/thread
>
> I revised the thread, saw that Uri made some modifications to the patch
> made by Supriya, but couldn't find a reference to this on the git log,
> so I am not sure if we already have this integrated.
>
> ? Needs rework ?
>
> --------------------------------------------------------------------------------
> Yolkfull Chow (yzhou@redhat.com)
>
> 1) fix a little problem for kvm_config.py
>
> Bugfix, easy to apply. Shouldn't require a lot of rework.
>
> Low hanging fruit
>
> --------------------------------------------------------------------------------
> Mike Burns (mburns@redhat.com)
>
> 1) Make code to remove kvm modules more robust
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/20/5787593/thread
>
> I have re-worked this patch from Mike and proposed upstream
>
> http://patchwork.test.kernel.org/patch/650/
>
> Since it's taking a bit longer than expected to review by upstream (will
> poke Martin), I have put it under kvm_utils, on my install refactor
> patch just for precaution. Once it gets into utils, I am going to drop
> the code on kvm_utils. Bottom line, I have assimilated this patch on
> other patches of mine (with Mike's permission of course).
>
> Assimilated into other patches
>
> --------------------------------------------------------------------------------
> Alexey Eromenko (aeromenk@redhat.com)
>
> 1) New test module: iperf
>
> http://kerneltrap.org/mailarchive/linux-kvm/2009/5/31/5840973/thread
>
> Needs rework
>
> --------------------------------------------------------------------------------
> Jason Wang (jasowang@redhat.com)
>
> 1) TAP network support in kvm-autotest
>
> I didn't find this one on the kernel trap archives
>
> Needs rework
>
>
>    

Re: [KVM-AUTOTEST] Patch queue for kvm autotest

[David Huff]

Lucas Meneghel Rodrigues wrote:
> Patch queue:
> 
> --------------------------------------------------------------------------------

Left out: [PATCH] Added functionality to the preprocessor to run scripts

This patch adds support for pre_command and post_command parameters in 
the congif file that specify a script to execute before/after a test is 
run to set the test environment.

I think it is important to implement this early so any additional tests 
or reworks can utilize this functionality.

I would mark this a pretty low hanging fruit, however there is an 
ongoing  discussion about how these scripts should be run and how the 
errors should be caught...

-D