Re: randconfig build error with next-20150113, in sound/pci/hda/patch_realtek.c

[Takashi Iwai <tiwai@suse.de>]

At Tue, 13 Jan 2015 11:39:12 -0700,
Jim Davis wrote:
> 
> Building with the attached random configuration file,
> 
> sound/built-in.o: In function `alc280_fixup_hp_gpio2_mic_hotkey':
> /home/jim/linux/sound/pci/hda/patch_realtek.c:3512: undefined
> reference to `input_allocate_device'
> /home/jim/linux/sound/pci/hda/patch_realtek.c:3520: undefined
> reference to `input_register_device'
> /home/jim/linux/sound/pci/hda/patch_realtek.c:3522: undefined
> reference to `input_free_device'
> /home/jim/linux/sound/pci/hda/patch_realtek.c:3550: undefined
> reference to `input_unregister_device'
> sound/built-in.o: In function `gpio2_mic_hotkey_event':
> /home/jim/linux/include/linux/input.h:392: undefined reference to `input_event'
> /home/jim/linux/include/linux/input.h:417: undefined reference to `input_event'
> /home/jim/linux/include/linux/input.h:392: undefined reference to `input_event'
> /home/jim/linux/include/linux/input.h:417: undefined reference to `input_event'
> make: *** [vmlinux] Error 1

Thanks.  So this is the case where the audio driver is built-in while
the input driver is module.

I guess the cleanest workaround is just to select INPUT from
SND_HDA_CODEC_REALTEK in sound/pci/hda/Kconfig.  Practically seen,
CONFIG_INPUT is always set, so there must be no real drawback with it.
(And we can reduce a few lines as a bonus.)

David, care to prepare a patch?


thanks,

Takashi