[systemd-devel] Status multiseat?

[Lennart Poettering]

On Mon, 05.12.11 21:02, Stef Bon (stefbon at gmail.com) wrote:

> Hi,
> 
> I'm informing the status of multiseat.
> 
> I've read here:
> 
> http://fedoraproject.org/wiki/Features/ckremoval
> 
> that there is worked on automatic detection of multiseat.
> 
> Some notes:
> 
> . automatic is only possible for local seats

Well, systemd's multi-seat support is exclusively for local seats. In
fact, the word "seat" for us implies "local".

> . maximum local seats is the minimum of terminals and keyboards (=N)

Well, the automatic mode means that a seat needs at least one display to
be around. A keyboard/mouse is optional.

> . part of N is "X" seat, where a mouse (or any other pointing device)
> is available and maybe required is a framebuffer with some
> capabilities.

Only the framebuffer needs to be available for a seat, everything else
is optional.

> . manual configuration overrides automatic detection. I've got two
> terminals in dual monitor (onescreen over two terminals) setup.

We can spawn seats only on invidividual framebuffers. framebuffers
cannot be shared between seats.

> . remote seat detection is never automatic, but happens on the fly
> when someone logs in.

remote logins are in fact completely orthogonal to systemd-logind, and
handled as virtual seat-less sessions, as is SSH and so on.

Lennart

-- 
Lennart Poettering - Red Hat, Inc.