support for HW_SKIP_CONSOLE breaks use by blind people

[Samuel Thibault]

Michal Suchanek, le Thu 10 Jan 2013 00:45:41 +0100, a écrit :
> Just backing out HW_SKIP_CONSOLE after years of dummy driver using it

Did I ever propose that?  (As a real upstream change I mean, not the
workaround I've provided in the Debian report which was never meant to
be applied upstream)

Also, backing out decades of dummy driver being useful for blind users
who have no screen shouldn't have been an option :)

> a proper fix needs something like:
> 
> 1) introduce a HW_OPEN_CONSOLE flag which is inverse of the HW_SKIP_CONSOLE flag
> 2) have all drivers but void and dummy set that - dummy has nice
> example of setting a flag
> 3) have the input hotplug code handle the flag in addition to the init code

That's more or less precisely what was suggested earlier in the thread,
and my precise enquiry waaaay earlier: which way should input drivers
set the flag, since the mechanism used for video drivers does not exist
for input drivers.

That said, it's of course way simpler, instead of having to modify all
drivers, to just make video-dummy and input-void set the SKIP flag,
assuming the default for a given driver is cleared (usual case), but the
default global being skip, as it is already ATM, so that allocation
happens as soon as a driver does not set the SKIP flag (which is
equivalent to what you propose, except not requiring modifying all
drivers).

Now, Peter also proposed to come back to the default behavior of
allocating a VT, unless a -novt option is passed. Just like we have
-novtswitch, -sharevts, vtXX. That looks much more coherent to me.

Samuel