[RFC v2 wayland-protocols] inputfd - direct input access protocol

[Peter Hutterer]

On Tue, Apr 25, 2017 at 11:34:06AM +0200, Bastien Nocera wrote:
> On Sun, 2017-04-23 at 09:47 -0700, Roderick Colenbrander wrote:
> > 
> <snip>
> > We originally discussed maybe using pipes and now more unix domain
> > sockets for 'non kernel evdev fds'. As part of the discussion we
> > realized ioctls wouldn't be available. Thinking about it some more, I
> > think the lack of ioctl support is problematic. Sure we can use fds
> > to
> > send 'struct input_event' over, but EVIOCGABS is critical for
> > determining capabilties and value ranges. Many others are useful too
> > like the revoke ones. How should this be dealt with?
> 
> To be honest, I expected the single file descriptor to work both ways,
> and with ioctls, but that's obviously not possible when the ioctl might
> be blocking.
> 
> I've looked through SDL, and it only uses ioctl()s for this initial
> state and for force feedback ("haptic"). A separate API for that looks
> unfortunately in order.

fwiw, the kernel has SYN_CONFIG but it's not well defined (or defined at
all) so right now the kernel pretty much guarantees that a device never
changes. And even if it didn't, too much of the userspace already relies on
that so it cannot be changed. Hence the ioctls only being used during
initialisation.

Cheers,
   Peter