Currently, the i2c and spi subdev creation helpers
are calling BUG() when passed a NULL v4l2_device parameter.

This makes little sense; simply returning NULL seems more
sensible.

These two helpers may already return NULL on error, so callers
should already be checking for this, or at least be prepared
for a NULL result.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>