virtio — archive

[Stefan Hajnoczi]

On Fri, Oct 30, 2020 at 01:16:45PM +0100, Cornelia Huck wrote: Let's clarify whether it is acceptable to keep the device without a spec indefinitely. I am starting to get the feeling people would like to define one-off VIRTIO devices that are unlikely to be implemented elsewhere. As VIRTIO expands we may find it acceptable to allow these use cases, but expectations should be clear: 1. VIRTIO spec features won't be added for out-of-tree devices. The user needs to be in the spec so that the design can be reviewed. 2. Out-of-tree devices that violate the VIRTIO device model may cease to work or be impossible to implement as the VIRTIO spec evolves. > Perhaps there needs to be some kind of annotation? If you reserve an ID > via the quick mechanism, it gets a "reserved until YYYY-MM-DD", and > that tag gets removed when a proper spec is merged? For > "in-the-wild-do-not-reuse" IDs, we'd still require a vote, as the > default should be to have a proper spec if you want something outside > of the private namespace. I'm not sure revoking an ID is realistic in the case where the owner refuses to upstream a spec but is shipping devices. The upstream device that reused the ID will suffer when collisions occur. Stefan