On Fri, 18 Oct 2019 15:47:39 +0200 Philipp Zabel <[email protected]> wrote:
> On Fri, 2019-10-18 at 08:51 +0000, Simon Ser wrote: > > On Thursday, October 17, 2019 11:15 PM, Drew DeVault <[email protected]> wrote: > > > > > On Thu Oct 17, 2019 at 6:08 PM Simon Ser wrote: > > > > > > > Should we keep the connector event, now that we have an unprivileged > > > > DRM FD? > > > > Alternatives include: > > > > > > > > - Let the client use the FD to get what it needs (EDID/a configured > > > > output/something else). > > > > > > > > > > Isn't this: > > > > > > > - Keep an event to advertise lease-able connector IDs > > > > > > What it's still there for? I'm not sure I understand, otherwise how is > > > the client supposed to know which DRM resources it can request a lease > > > for? How does it encode those intentions into a lease request? afaict > > > the connector is still necessary. > > > > The client can recognize EDIDs it's interested in, and then try to > > lease these connectors. The lease will fail if the connectors aren't > > leasable. This isn't a big deal for e.g. VR and we can always add the > > connector event later if needed. > > If there was no event for the appearance of leasable connectors, the > client would have to listen to hotplug uevents itself if it wants to be > capable of handling HMDs plugged in later. Is this something we want to > require? Supporting hotplug is a good question. I didn't think the connector events were for that, though. I think hotplug and hot-unplug should be discussed in the spec, how they are expected to work. DRM hotplug events are udev events which (for now) just say "hey, go re-scan stuff, something probably changed". I don't think clients should need to listen to udev events as well. Thanks, pq
pgpF55VUk7wOD.pgp
Description: OpenPGP digital signature
_______________________________________________ wayland-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/wayland-devel
