ASoC: rockchip: rk3288_hdmi_analog: consider CPU-Platform possibility

commit 9c21e82c ("ASoC: rockchip: rk3288_hdmi_analog: don't select
unnecessary Platform")

Current ALSA SoC avoid to add duplicate component to rtd,
and this driver was selecting CPU component as Platform component.
Thus, above patch removed Platform settings from this driver,
because it assumed these are same component.

But, some CPU driver is using generic DMAEngine, in such case, both
CPU component and Platform component will have same of_node/name.
In other words, there are some components which are different but
have same of_node/name.

In such case, Card driver definitely need to select Platform even
though it is same as CPU.
It is depends on CPU driver, but is difficult to know it from Card driver.
This patch reverts above patch.

Fixes: commit 9c21e82c ("ASoC: rockchip: rk3288_hdmi_analog: don't select unnecessary Platform")
Signed-off-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

sound/soc/rockchip/rk3288_hdmi_analog.c

@@ -142,7 +142,8 @@ static const struct snd_soc_ops rk_ops = {
 SND_SOC_DAILINK_DEFS(audio,
 	DAILINK_COMP_ARRAY(COMP_EMPTY()),
 	DAILINK_COMP_ARRAY(COMP_CODEC(NULL, NULL),
-			   COMP_CODEC("hdmi-audio-codec.2.auto", "i2s-hifi")));
+			   COMP_CODEC("hdmi-audio-codec.2.auto", "i2s-hifi")),
+	DAILINK_COMP_ARRAY(COMP_EMPTY()));
 
 static struct snd_soc_dai_link rk_dailink = {
 	.name = "Codecs",
@@ -236,6 +237,8 @@ static int snd_rk_mc_probe(struct platform_device *pdev)
 		return -EINVAL;
 	}
 
+	rk_dailink.platforms->of_node = rk_dailink.cpus->of_node;
+
 	ret = snd_soc_of_parse_audio_routing(card, "rockchip,routing");
 	if (ret) {
 		dev_err(&pdev->dev,