Re: fundamentally broken on powerpc
Martin Michlmayr wrote:
>
> > On PowerMacs, console-data blindly assumes that the keyboard sends ADB
> > keycodes and only offers such keymaps. If
> > /proc/sys/dev/mac_hid/keyboard_sends_linux_keycodes contains 1 (or
> > doesn't exist), an i386 keymap is needed though, or the keyboard is
> > unusable.
>
> I've done an NMU which should fix this. I'd be glad if you could test
> the package in sid and let me know if it works for you. My question
> is whether this applies to all powerpc sub-architectures or only to
> powerpc/mac. console-common knows the following ones:
>
> 'powerpc/amiga' => [ 'amiga' ],
> 'powerpc/chrp' => [ 'pc', 'mac' ],
> 'powerpc/mac' => [ 'mac' ],
> + 'powerpc/macpc' => [ 'pc' ],
> 'powerpc/prep' => [ 'pc' ],
>
> The macpc was introduced by me and is used when
> keyboard_sends_linux_keycodes is 1 or doesn't exist.
Doesn't sound bad, but apparently it still doesn't work right if
/proc/sys/dev/mac_hid/keyboard_sends_linux_keycodes is missing. From the patch
you sent me privately, I don't understand why though.
> > If /proc/sys/dev/mac_hid/keyboard_sends_linux_keycodes contains 0
> > OTOH, the user should be strongly encouraged to disable ADB keycodes in
> > the kernel (like it will probably be done in the default woody kernel),
> > an i386 keymap should be installed and the sysctl should be set to 1.
>
> So basically this means that I use an i386 keymap in _all cases_, right?
> (Note that current console-* ignores this wish and only fixes the
> critical bug (i.e. what is described in the first paragraph)).
I and Ethan Benson agreed on IRC that it is probably better to stop supporting
mac keymaps altogether.
What about this: Change the mac line to 'powerpc/mac' => [ 'pc' ] and alert
the user if /proc/sys/dev/mac_hid/keyboard_sends_linux_keycodes contains 0
(and set it to 1)?
Looking forward to feedback from the powerpc list...
--
Earthling Michel Dänzer (MrCooper) \ Debian GNU/Linux (powerpc) developer
CS student, Free Software enthusiast \ XFree86 and DRI project member
Reply to: