On Thu, 2022-04-28 at 13:02 +0100, Mark Brown wrote:
On Thu, Apr 28, 2022 at 05:33:50PM +0800, Jiaxin Yu wrote:
Add mt8186 platform and affiliated driver.
Signed-off-by: Jiaxin Yu jiaxin.yu@mediatek.com
sound/soc/mediatek/Kconfig | 44 + sound/soc/mediatek/Makefile | 1 + sound/soc/mediatek/mt8186/Makefile | 22 + sound/soc/mediatek/mt8186/mt8186-afe-common.h | 235 ++ .../soc/mediatek/mt8186/mt8186-afe-control.c | 261 ++ sound/soc/mediatek/mt8186/mt8186-afe-pcm.c | 3005 +++++++++++++++++ .../mediatek/mt8186/mt8186-interconnection.h | 69 + .../soc/mediatek/mt8186/mt8186-misc-control.c | 294 ++ .../mediatek/mt8186/mt8186-mt6366-common.c | 59 + .../mediatek/mt8186/mt8186-mt6366-common.h | 17 + sound/soc/mediatek/mt8186/mt8186-reg.h | 2913 ++++++++++++++++ 11 files changed, 6920 insertions(+)
This looks mostly good though it is enormous so I might've missed some things. The patch series is already very large but it might still be worth splitting this up a bit more, perhaps split the code and data tables/register definitions into separate patches?
Yes, agree with you.
I will spit them into three patches:
PATCH 1: - mt8186-reg.h - mt8186-interconnection.h - mt8186-misc-control.c
PATCH 2: - mt8186-mt6366-common.c - mt8186-mt6366-common.h
PATCH 3: - sound/soc/mediatek/Kconfig - sound/soc/mediatek/Makefile - sound/soc/mediatek/mt8186/Makefile - sound/soc/mediatek/mt8186/mt8186-afe-common.h - .../soc/mediatek/mt8186/mt8186-afe-control.c - sound/soc/mediatek/mt8186/mt8186-afe-pcm.c
A few relatively minor issues with the controls.
+/* this order must match reg bit amp_div_ch1/2 */ +static const char * const mt8186_sgen_amp_str[] = {
- "1/128", "1/64", "1/32", "1/16", "1/8", "1/4", "1/2", "1" };
+static const char * const mt8186_sgen_mute_str[] = {
- "Off", "On"
+};
On/off controls should be normal Switch controls not enums so userspace can display things sensibly.
+static int mt8186_sgen_set(struct snd_kcontrol *kcontrol,
struct snd_ctl_elem_value *ucontrol)+{
- struct snd_soc_component *cmpnt =
snd_soc_kcontrol_component(kcontrol);
- struct mtk_base_afe *afe =
snd_soc_component_get_drvdata(cmpnt);
- struct mt8186_afe_private *afe_priv = afe->platform_priv;
- struct soc_enum *e = (struct soc_enum *)kcontrol-
private_value;
- int mode;
- int mode_idx;
- if (ucontrol->value.enumerated.item[0] >= e->items)
return -EINVAL;...
0x3f << INNER_LOOP_BACK_MODE_SFT);- }
- afe_priv->sgen_mode = mode;
- return 0;
+}
This should return 1 if the value is different from the previous value so event generation works, please run mixer-test against a system using the driver to help spot issues like this. The same issue applies to at least some of the other custom controls.
Got it.
+static int mt8186_sgen_mute_set(struct snd_kcontrol *kcontrol,
struct snd_ctl_elem_value *ucontrol)+{
- struct snd_soc_component *cmpnt =
snd_soc_kcontrol_component(kcontrol);
- struct mtk_base_afe *afe =
snd_soc_component_get_drvdata(cmpnt);
- struct soc_enum *e = (struct soc_enum *)kcontrol-
private_value;
- int mute;
- if (ucontrol->value.enumerated.item[0] >= e->items)
return -EINVAL;- mute = ucontrol->value.integer.value[0];
- dev_dbg(afe->dev, "%s(), kcontrol name %s, mute %d\n",
__func__, kcontrol->id.name, mute);- if (strcmp(kcontrol->id.name, SGEN_MUTE_CH1_KCONTROL_NAME) ==
- {
regmap_update_bits(afe->regmap, AFE_SINEGEN_CON0,MUTE_SW_CH1_MASK_SFT,mute << MUTE_SW_CH1_SFT);- } else {
regmap_update_bits(afe->regmap, AFE_SINEGEN_CON0,MUTE_SW_CH2_MASK_SFT,mute << MUTE_SW_CH2_SFT);- }
- return 0;
+}
I can't tell why some of these are done with custom code rather than using a normal SOC_SINGLE()?
Yes, it's better to use SOC_SINGLE. I will fix them in next version.
Thanks, Jiaxin.yu