On Wed, Nov 20, 2019 at 05:05:00PM -0500, Michael S. Tsirkin wrote: > On Wed, Nov 20, 2019 at 12:45:25PM -0400, Jason Gunthorpe wrote: > > > > For instance, this VFIO based approach might be very suitable to the > > > > intel VF based ICF driver, but we don't yet have an example of non-VF > > > > HW that might not be well suited to VFIO. > > > > > > I don't think we should keep moving the goalposts like this. > > > > It is ABI, it should be done as best we can as we have to live with it > > for a long time. Right now HW is just starting to come to market with > > VDPA and it feels rushed to design a whole subsystem style ABI around > > one, quite simplistic, driver example. > > Well one has to enable hardware in some way. It's not really reasonable > to ask for multiple devices to be available just so there's a driver and > people can use them. Er, this has actually been a fairly standard ask for new subsystems. I think virtio is well grounded here compared to other things I've seen, but it should still be done with a lot more NIC community involvement. > At this rate no one will want to be the first to ship new devices ;) Why? > > > If people write drivers and find some infrastruture useful, > > > and it looks more or less generic on the outset, then I don't > > > see why it's a bad idea to merge it. > > > > Because it is userspace ABI, caution is always justified when defining > > new ABI. > > Reasonable caution, sure. Asking Alex to block Intel's driver until > someone else catches up and ships competing hardware isn't reasonable > though. If that's your proposal I guess we'll have to agree to disagree. Vendors may be willing to participate, as Mellanox is doing, pre-product. Jason
