[bug report] ASoC: mediatek: mt8186_mt6366_rt1019_rt5682s: add rt5650 support

[bug report] ASoC: mediatek: mt8186_mt6366_rt1019_rt5682s: add rt5650 support

[Dan Carpenter]

Hello xiazhengqiao,

The patch d88c43383101: "ASoC: mediatek:
mt8186_mt6366_rt1019_rt5682s: add rt5650 support" from Oct 19, 2023
(linux-next), leads to the following Smatch static checker warning:

	sound/soc/mediatek/mt8186/mt8186-mt6366-rt1019-rt5682s.c:198 mt8186_rt5682s_init()
	warn: does endianness matter for 'type'?

sound/soc/mediatek/mt8186/mt8186-mt6366-rt1019-rt5682s.c
    161 static int mt8186_rt5682s_init(struct snd_soc_pcm_runtime *rtd)
    162 {
    163         struct snd_soc_component *cmpnt_afe =
    164                 snd_soc_rtdcom_lookup(rtd, AFE_PCM_NAME);
    165         struct mtk_base_afe *afe = snd_soc_component_get_drvdata(cmpnt_afe);
    166         struct mtk_soc_card_data *soc_card_data =
    167                 snd_soc_card_get_drvdata(rtd->card);
    168         struct mt8186_mt6366_rt1019_rt5682s_priv *priv = soc_card_data->mach_priv;
    169         struct snd_soc_jack *jack = &priv->headset_jack;
    170         struct snd_soc_component *cmpnt_codec =
    171                 snd_soc_rtd_to_codec(rtd, 0)->component;
    172         int ret;
    173         int type;
    174 
    175         ret = mt8186_dai_i2s_set_share(afe, "I2S1", "I2S0");
    176         if (ret) {
    177                 dev_err(rtd->dev, "Failed to set up shared clocks\n");
    178                 return ret;
    179         }
    180 
    181         ret = snd_soc_card_jack_new_pins(rtd->card, "Headset Jack",
    182                                     SND_JACK_HEADSET | SND_JACK_BTN_0 |
    183                                     SND_JACK_BTN_1 | SND_JACK_BTN_2 |
    184                                     SND_JACK_BTN_3,
    185                                     jack, mt8186_jack_pins,
    186                                     ARRAY_SIZE(mt8186_jack_pins));
    187         if (ret) {
    188                 dev_err(rtd->dev, "Headset Jack creation failed: %d\n", ret);
    189                 return ret;
    190         }
    191 
    192         snd_jack_set_key(jack->jack, SND_JACK_BTN_0, KEY_PLAYPAUSE);
    193         snd_jack_set_key(jack->jack, SND_JACK_BTN_1, KEY_VOICECOMMAND);
    194         snd_jack_set_key(jack->jack, SND_JACK_BTN_2, KEY_VOLUMEUP);
    195         snd_jack_set_key(jack->jack, SND_JACK_BTN_3, KEY_VOLUMEDOWN);
    196 
    197         type = SND_JACK_HEADSET | SND_JACK_BTN_0 | SND_JACK_BTN_1 | SND_JACK_BTN_2 | SND_JACK_BTN_3;
--> 198         return snd_soc_component_set_jack(cmpnt_codec, jack, (void *)&type);

This is an unpublished Smatch check where I manually review casts to see
if they are correct.  Quite often they aren't because of an endian bug
or a 64 bit vs 32 bit issue.

Here it's not clear to me what's happening.  Normally with this sort of
pass a void pointer code, you can tie it very easily to the same driver.
But in this case it's much more difficult.

There are two functions which use the void *data pointer,
rt5640_set_jack() and rt5645_component_set_jack().  One takes an int and
the other takes a struct rt5640_set_jack_data pointer.  So presumably
we know that the cmpnt_codec->driver->set_jack points to
rt5645_component_set_jack().  But how do we know that?

Is there a trick for me as a reviewer to use?

    199 }

regards,
dan carpenter

Re: [bug report] ASoC: mediatek: mt8186_mt6366_rt1019_rt5682s: add rt5650 support

[Dan Carpenter]

Thanks so much.  This is usefull.

regards,
dan carpenter