On Thu, Oct 16, 2025 at 08:19:37PM +0100, Daniel P. Berrangé wrote: > On Thu, Oct 16, 2025 at 07:51:42PM +0100, Daniel P. Berrangé wrote: > > On Thu, Oct 16, 2025 at 02:40:58PM -0400, Peter Xu wrote: > > > On Thu, Oct 16, 2025 at 12:23:35PM +0300, Vladimir Sementsov-Ogievskiy > > > wrote: > > > > On 16.10.25 11:32, Daniel P. Berrangé wrote: > > > > > On Thu, Oct 16, 2025 at 12:02:45AM +0300, Vladimir > > > > > Sementsov-Ogievskiy wrote: > > > > > > 1. Remote migration: we can't reuse backends (files, sockets, host > > > > > > devices), as > > > > > > we are moving to another host. So, we don't enable > > > > > > "backend-transfer". We don't > > > > > > transfer the backend, we have to initialize new backend on another > > > > > > host. > > > > > > > > > > > > 2. Local migration to update QEMU, with minimal freeze-time and > > > > > > minimal > > > > > > extra actions: use "backend-transfer", exactly to keep the backends > > > > > > (vhost-user-server, TAP device in kernel, in-kernel vfio device > > > > > > state, etc) > > > > > > as is. > > > > > > > > > > > > 3. Local migration, but we want to reconfigure some backend, or > > > > > > switch > > > > > > to another backend. We disable "backend-transfer" for one device. > > > > > > > > > > This implies that you're changing 'backend-transfer' against the > > > > > device at time of each migration. > > > > > > > > > > This takes us back to the situation we've had historically where the > > > > > behaviour of migration depends on global properties the mgmt app has > > > > > set prior to the 'migrate' command being run. We've just tried to get > > > > > away from that model by passing everything as parameters to the > > > > > migrate command, so I'm loathe to see us invent a new way to have > > > > > global state properties changing migration behaviour. > > > > > > > > > > This 'backend-transfer' device property is not really a device > > > > > property, > > > > > it is an indirect parameter to the 'migrate' command. > > > > > > I was not seeing it like that. > > > > > > I was treating per-device parameter to be a flag showing whether the > > > device > > > is capable of passing over FDs, which is more like a device attribute. > > Whether a backend is technically capable of transfer shouldn't require a > user specified property - there should be an internal API to query whether > the current backend configuration is transferrable or not, based on the > code implementation. Allowing a mgmt app to specify this can only lead > to mistakes, because they don't know the internal constraints of the > implementation. > > The mgmt app should only be concerned with whether they want to transfer > a backend or not which is a time-of-use decision rather than launch time > decision.
IMHO the per-device property, when available, should always mean it fully support the feature, when it is turned ON. I also think above statement matches exactly how I see it.. I never expected mgmt to toggle the per-device properties, as I just left similar statements in another reply. That's also why I think the global backend-transfer should be the only thing exposed to mgmt. So even if the device properties would exist, they should only be used in compat properties for the upstream QEMUs. They're still needed, and be helpful when other devices introduce some similar concepts to support fd passover, then on some machine types when the global feature enabled, QEMU will automatically do fd-pass for some devices and some not, based on the machine type. Thanks, -- Peter Xu
