qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v6 0/8] migration: Add switchover ack capability and VFIO pre


From: Avihai Horon
Subject: Re: [PATCH v6 0/8] migration: Add switchover ack capability and VFIO precopy support
Date: Thu, 22 Jun 2023 17:19:58 +0300
User-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101 Thunderbird/102.12.0


On 22/06/2023 17:02, Cédric Le Goater wrote:
External email: Use caution opening links or attachments


On 6/22/23 14:16, Avihai Horon wrote:

On 22/06/2023 14:34, Cédric Le Goater wrote:
External email: Use caution opening links or attachments


On 6/21/23 21:28, Alex Williamson wrote:
On Wed, 21 Jun 2023 14:11:53 +0300
Avihai Horon <avihaih@nvidia.com> wrote:

Hello everyone,

The latest changes to migration qtest made the tests run non-live by
default. I am posting this v6 to change back the switchover-ack
migration test to run live as it used to (because the source VM needs to
be running to consider the switchover ACK when deciding to do the
switchover or not).

Changes from v5 [7]:
* Rebased on latest master branch.
* Made switchover-ack migration test run live again (I kept the R-bs as
   this was the original behavior when they were given).
* Dropped patch #8 (x-allow-pre-copy property). (Alex)
* Adjusted patch #9 commit message according to drop of patch #8.
* Added R-b to patch #9 and Tested-by tags to the series.

I think Cédric is going to handle the pull request for this, so...

Yes. I have a branch ready with this series and a few other fixes. Unless
Juan want to handle it.


Avihai,

QEMU seems ready to allow migration when support is available in the
kernel now and so, we can set "x-enable-migration" to true by default.
Are you OK with that ?

Yes, sure.
But as per previous off-list discussions, it's not merely changing x-enable-migration to true by default, right?
If I recall correctly, Alex wanted to remove the x- prefix,

OK. We have to start one day :)

change it to ON_OFF_AUTO and some other stuff.

Is it to silently disable migration support if some failure occurs
at the kernel level when ON_OFF_AUTO_AUTO is set and report an error
when ON_OFF_AUTO_ON is set ?

Yes, or other errors.
And also to require device dirty tracking with AUTO.



Do you have some patches ready for that?

nope.

Or you want me to do it?

If you have time, yes please.

Sure, so I will prepare the patches and send them.

Thanks.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]