usb/144414: Apple "Fn" key doesn't work properly

[Steven Noonan]

On Sun, Mar 14, 2010 at 7:23 AM, Hans Petter Selasky <hselasky at c2i.net> wrote:
> On Sunday 14 March 2010 15:49:37 Steven Noonan wrote:
>> On Sun, Mar 14, 2010 at 6:46 AM, Kai Wang <kaiwang27 at gmail.com> wrote:
>> > On Sun, Mar 14, 2010 at 06:37:04AM -0800, Steven Noonan wrote:
>> >> On Sun, Mar 14, 2010 at 6:27 AM, Hans Petter Selasky <hselasky at c2i.net>
> wrote:
>> >> > On Sunday 14 March 2010 15:01:39 Steven Noonan wrote:
>> >> >> On Sun, Mar 14, 2010 at 5:48 AM, Hans Petter Selasky
>> >> >> <hselasky at c2i.net>
>> >> >
>> >> > wrote:
>> >> >> > On Sunday 14 March 2010 14:22:33 Steven Noonan wrote:
>> >> >> >> On Sun, Mar 14, 2010 at 3:01 AM, Hans Petter Selasky
>> >> >> >> <hselasky at c2i.net>
>> >> >> >
>> >> >> > wrote:
>> >> >> >> > On Sunday 14 March 2010 11:30:04 Steven Noonan wrote:
>> >> >> >> >> The following reply was made to PR usb/144414; it has been
>> >> >> >> >> noted by GNATS.
>> >> >> >> >>
>> >> >> >> >> From: Steven Noonan <steven at uplinklabs.net>
>> >> >> >> >> To: perryh at pluto.rain.com
>> >> >> >> >> Cc: freebsd-usb at freebsd.org, freebsd-gnats-submit at freebsd.org
>> >> >> >> >> Subject: Re: usb/144414: Apple "Fn" key doesn't work properly
>> >> >> >> >> Date: Sun, 14 Mar 2010 03:26:19 -0700
>> >> >> >> >>
>> >> >> >> >>  On Thu, Mar 4, 2010 at 12:01 AM, Steven Noonan
>> >> >> >> >> <steven at uplinklabs.net> wrot=
>> >> >> >> >>
>> >> >> >> >>  e:
>> >> >> >> >>  > On Wed, Mar 3, 2010 at 11:00 PM, Steven Noonan
>> >> >> >> >> <steven at uplinklabs.net> > wr=
>> >> >> >> >>
>> >> >> >> >>  ote:
>> >> >> >> >>  >> On Wed, Mar 3, 2010 at 10:36 PM,
>> >> >> >> >> =C2=A0<perryh at pluto.rain.com> wrote: >>> Steven Noonan
>> >> >> >> >> <steven at uplinklabs.net> wrote: >>>> Interestingly, my tilde key
>> >> >> >> >> doesn't work either (though the key >>>> press is detected, no
>> >> >> >> >> character shows when the key is pressed). >>>
>> >> >> >> >>  >>> Any chance it is configured as a "dead" key? =C2=A0If you
>> >> >> >> >> press >>> tilde followed by n do you get an n with a tilde over
>> >> >> >> >> it? >>
>> >> >> >> >>  >> Nope. It simply does nothing in the console.
>> >> >> >> >>  >>
>> >> >> >> >>  >> But in X11, it does something very odd. Shift+Tilde Key
>> >> >> >> >> gives me '>', >> and Tilde Key gives me '<'. And Alt+Tilde
>> >> >> >> >> gives me... What? I don't >> even know what action. It's
>> >> >> >> >> grabbing some arbitrary command in my >> .bash_history. It gave
>> >> >> >> >> me the first item in my .bash_history the first >> time I tried
>> >> >> >> >> it. Then I tried an arbitrary command ("echo"), and then >>
>> >> >> >> >> Alt+Tilde gave me the second command in my .bash_history. Whaa?
>> >> >> >> >>
>> >> >> >> >>  Any more ideas/news on this from anyone?
>> >> >> >> >>
>> >> >> >> >>  The tilde key thing is especially irritating.
>> >> >> >> >
>> >> >> >> > static uint8_t
>> >> >> >> > ukbd_apple_swap(uint8_t keycode) {
>> >> >> >> >        switch (keycode) {
>> >> >> >> >        case 0x35: return 0x64;
>> >> >> >> >        case 0x64: return 0x35;
>> >> >> >> >        default: return keycode;
>> >> >> >> >        }
>> >> >> >> > }
>> >> >> >> >
>> >> >> >> > Can you try to change the function above in ukbd.c to only
>> >> >> >> > return keycode? Any difference?
>> >> >> >>
>> >> >> >> Yep, commenting the two case lines brought my tilde key back. Any
>> >> >> >> idea what the ukbd_apple_swap() function was _supposed_ to be
>> >> >> >> doing?
>> >> >> >>
>> >> >> >> Also, there's only one thing left (input-wise) that I can't get to
>> >> >> >> work, and that's the 'delete' key (Fn+Backspace). I'm pretty
>> >> >> >> certain that it's not a problem with Fn+Backspace failing to map
>> >> >> >> to Delete, because my non-Mac i386 box has the same problem.
>> >> >> >>
>> >> >> >> Everything I read online keeps saying to muck with an .inputrc
>> >> >> >> file (for bash, anyway), but I haven't been able to get it to
>> >> >> >> work. Places I've looked have said to add this to .inputrc:
>> >> >> >>
>> >> >> >> "\e[3~": delete-char
>> >> >> >>
>> >> >> >> But this seems to have no effect (and I tried bind -f .inputrc in
>> >> >> >> case my INPUTRC environment variable wasn't working). Is there
>> >> >> >> something FreeBSD-specific I'm not seeing, or what?
>> >> >> >>
>> >> >> >> - Steven
>> >> >> >
>> >> >> > Hi Steven,
>> >> >> >
>> >> >> > Can you talk this over with Christoph and send me a patch when you
>> >> >> > agreed? He's the one that made the Apple-SWAP key patch. Meanwhile
>> >> >> > I suggest the following patch:
>> >> >> >
>> >> >> > ==== //depot/projects/usb/src/sys/dev/usb/input/ukbd.c#47 -
>> >> >> > src/sys/dev/usb/input/ukbd.c ====
>> >> >> > @@ -896,8 +896,7 @@
>> >> >> >                    hid_input, 0, &sc->sc_loc_apple_fn, &flags,
>> >> >> >                    &temp_id)) {
>> >> >> >                        if (flags & HIO_VARIABLE)
>> >> >> > -                               sc->sc_flags |= UKBD_FLAG_APPLE_FN
>> >> >> > | -                                   UKBD_FLAG_APPLE_SWAP;
>> >> >> > +                               sc->sc_flags |= UKBD_FLAG_APPLE_FN;
>> >> >> >                        DPRINTFN(1, "Found Apple FN-key\n");
>> >> >> >                        apple_keys = 1;
>> >> >> >                        sc->sc_kbd_id = temp_id;
>> >> >> >
>> >> >> >
>> >> >> > The hardware that Christoph was using had an Eject-key, while yours
>> >> >> > didn't. I think the patch above will make both cases work.
>> >> >> >
>> >> >> > Christoph, do you have any comments?
>> >> >> >
>> >> >> > http://p4db.freebsd.org/fileViewer.cgi?FSPC=//depot/projects/usb/sr
>> >> >> >c/sys/ dev/usb/input/ukbd.c&REV=47
>> >> >>
>> >> >> Not quite true, my keyboard _does_ have an Eject key. The code
>> >> >> doesn't detect it though.
>> >> >>
>> >> >> - Steven
>> >> >
>> >> > Could you try to figure out what key-press number corresponds to the
>> >> > eject key?
>> >> >
>> >> > --HPS
>> >>
>> >> I'm not certain how to do so. It's somewhat of a chicken-and-the-egg
>> >> problem: unless it detects that the key is there, I can't find it.
>> >> Enabling 'hw.usb.ukbd.debug' doesn't print anything when Eject is
>> >> pressed.
>> >
>> > It's very possible that your Eject key is in a separate interface and
>> > is handled by uhid(4) instead of ukbd(4).
>> >
>> > /Kai
>>
>> Enabling 'hw.usb.uhid.debug' doesn't print anything on an Eject
>> keypress either, but that could just mean that uhid(4) doesn't have
>> necessary DPRINTFs to show it.
>>
>> - Steven
>>
>
> You need to open /dev/uhidX before it will print anything.
>
> --HPS
>

Aha. Doing 'cat /dev/uhid0 > /dev/null' and then hitting eject yielded:

"uhid_intr_callback: transferred!"

in dmesg.

- Steven