RE: [PATCH v2 2/2] vfio/migration: Make VFIO migration non-experimental

[Duan, Zhenzhong]

>-----Original Message-----
>From: Cédric Le Goater <clg@redhat.com>
>Sent: Thursday, June 29, 2023 12:04 AM
>To: Martins, Joao <joao.m.martins@oracle.com>; Avihai Horon
><avihaih@nvidia.com>
>Cc: Alex Williamson <alex.williamson@redhat.com>; Juan Quintela
><quintela@redhat.com>; Peter Xu <peterx@redhat.com>; Leonardo Bras
><leobras@redhat.com>; Duan, Zhenzhong <zhenzhong.duan@intel.com>;
>Yishai Hadas <yishaih@nvidia.com>; Jason Gunthorpe <jgg@nvidia.com>;
>Maor Gottlieb <maorg@nvidia.com>; Kirti Wankhede
><kwankhede@nvidia.com>; Tarun Gupta <targupta@nvidia.com>; qemu-
>devel@nongnu.org
>Subject: Re: [PATCH v2 2/2] vfio/migration: Make VFIO migration non-
>experimental
>
>On 6/28/23 16:51, Joao Martins wrote:
>> On 28/06/2023 13:54, Cédric Le Goater wrote:
>>> On 6/28/23 09:31, Avihai Horon wrote:
>>>> The major parts of VFIO migration are supported today in QEMU. This
>>>> includes basic VFIO migration, device dirty page tracking and
>>>> precopy support.
>>>>
>>>> Thus, at this point in time, it seems appropriate to make VFIO
>>>> migration
>>>> non-experimental: remove the x prefix from enable_migration
>>>> property, change it to ON_OFF_AUTO and let the default value be AUTO.
>>>>
>>>> In addition, make the following adjustments:
>>>> 1. When enable_migration is ON and migration is not supported, fail
>>>> VFIO
>>>>      device realization.
>>>> 2. When enable_migration is AUTO (i.e., not explicitly enabled),
>>>> require
>>>>      device dirty tracking support. This is because device dirty
>>>> tracking
>>>>      is currently the only method to do dirty page tracking, which
>>>> is
>>>>      essential for migrating in a reasonable downtime. Setting
>>>>      enable_migration to ON will not require device dirty tracking.
>>>> 3. Make migration error and blocker messages more elaborate.
>>>> 4. Remove error prints in vfio_migration_query_flags().
>>>> 5. Rename trace_vfio_migration_probe() to
>>>>      trace_vfio_migration_realize().
>>>>
>>>> Signed-off-by: Avihai Horon <avihaih@nvidia.com>
>>>
>>>
>>> We should rework the return value of most of the routines called by
>>> vfio_migration_realize() and simply use a bool. I think Zhenzhong is
>>> working it.
>>>
>>> Zhenzhong,
>>>
>>> When you resend v4 of the "VFIO migration related refactor and bug fix"
>>> series, please rebase on this patch since it should be merged.
>>>
>>
>> This, and his switchover-ack series from Avihai that preceeds it.
>>
>> Perhaps it might be easier to point to your tree:branch where you are
>> queueing all the patches?
>>
>
>Sure.
>
>I track QEMU patches for various subsystems under :
>
>  https://github.com/legoater/qemu
>
>
>VFIO candidates are under :
>
>   https://github.com/legoater/qemu/tree/vfio-8.1
>
>This is a wip tree, patches come and go. It contains the VFIO patches of the
>day/week, good for testing new ideas and checking CI.
>
>
>The vfio-next branch contains what I am 90% sure to send upstream :
>
>  https://github.com/legoater/qemu/tree/vfio-next

Great, I'll rebase on this branch, thanks for sharing.

Regards
Zhenzhong