[PATCH] do not feed keyboard device path in X devd backend

[Alex Kozlov]

On Thu, Feb 13, 2014 at 09:32:40PM +0000, Robert Millan wrote:
> On 13/02/2014 16:28, Alex Kozlov wrote:
> >> [  1002.547] (EE) Keyboard: cannot open "/dev/ukbd0"
> >> [  1002.547] (EE) PreInit returned 8 for "Keyboard"
> >> [  1002.547] (II) UnloadModule: "kbd"
> > This should be worked around by the next code:
> > 
> > 269 +    if (attrs.flags & ATTR_KEYBOARD) {
> > 270 +      /*
> > 271 +       * Don't pass device option if keyboard is attached to console (open fails),
> > 272 +       * thus activating special logic in xf86-input-keyboard.
> > 273 +       */
> > 274 +       fd = open(path, O_RDONLY | O_NONBLOCK | O_EXCL);
> > 275 +       if (fd > 0) {
> > 276 +          close(fd);
> > 277 +          options = input_option_new(options, "device", path);
> > 278 +       }
> > 279 +    } else {
> > 280 +          options = input_option_new(options, "device", path);
> > 281 +    }
> 
> Where's this code? I can't see it in x11-servers/xorg-server/files/extra-config_devd.c
http://trillian.chruetertee.ch/ports/browser/trunk/x11-servers/xorg-server/files/extra-devd
plus
http://trillian.chruetertee.ch/ports/browser/trunk/x11-servers/xorg-server/files/extra-new-config_devd
 
You need revision 1282 of x11-servers/xorg-server.

> >> [  1002.547] (II) config/devd: Adding input device Keyboard (/dev/atkbd0)
> >> [  1002.547] (**) Keyboard: Applying InputClass "kbd catchall"
> >> [  1002.547] (II) Using input driver 'kbd' for 'Keyboard'
> >> [  1002.547] (**) Keyboard: always reports core events
> >> [  1002.547] (**) Keyboard: always reports core events
> >> [  1002.547] (**) Option "Protocol" "standard"
> >> [  1002.547] (**) Option "Device" "/dev/atkbd0"
> >> [  1002.547] (EE) Keyboard: cannot open "/dev/atkbd0"
> >> [  1002.547] (EE) PreInit returned 8 for "Keyboard"
> >> [  1002.547] (II) UnloadModule: "kbd"
> > Btw, why both your keyboards are attached to console?
> I only have one keyboard (USB). I have no idea why does atkbd0 exist (but luckily
> its presence doesn't harm).
Hmm, I would like to see a log with up-to-date patch from xorg-dev.


-- 
Alex