On Tue, 19 Jan 2021 21:33:04 +0100,
Hans de Goede wrote:
>
> Hi,
>
> On 1/19/21 5:33 PM, Limonciello, Mario wrote:
> >>>> -----Original Message-----
> >>>> From: Yuan, Perry <Perry_Yuan@xxxxxxxx>
> >>>> Sent: Tuesday, January 12, 2021 11:18
> >>>> To: oder_chiou@xxxxxxxxxxx; perex@xxxxxxxx; tiwai@xxxxxxxx;
> >>>> hdegoede@xxxxxxxxxx; mgross@xxxxxxxxxxxxxxx
> >>>> Cc: lgirdwood@xxxxxxxxx; broonie@xxxxxxxxxx; alsa-devel@xxxxxxxxxxxxxxxx;
> >>>> linux-kernel@xxxxxxxxxxxxxxx; platform-driver-x86@xxxxxxxxxxxxxxx; Yuan,
> >>>> Perry; Limonciello, Mario
> >>>> Subject: [PATCH v3 3/3] ASoC: rt715:add micmute led state control supports
> >>>>
> >>>> From: Perry Yuan <perry_yuan@xxxxxxxx>
> >>>>
> >>>> Some new Dell system is going to support audio internal micphone
> >>>> privacy setting from hardware level with micmute led state changing
> >>>> When micmute hotkey pressed by user, soft mute will need to be enabled
> >>>> firstly in case of pop noise, and codec driver need to react to mic
> >>>> mute event to EC(embedded controller) notifying that SW mute is completed
> >>>> Then EC will do the hardware mute physically within the timeout reached
> >>>>
> >>>> This patch allow codec rt715 driver to ack EC when micmute key pressed
> >>>> through this micphone led control interface like hda_generic provided
> >>>> ACPI method defined in dell-privacy micmute led trigger will be called
> >>>> for notifying the EC that software mute has been completed
> >>>>
> >>>> Signed-off-by: Perry Yuan <perry_yuan@xxxxxxxx>
> >>>>
> >>>> --------
> >>>> v2 -> v3
> >>>> * simplify the patch to reuse some val value
> >>>> * add more detail to the commit info
> >>>>
> >>>> v1 -> v2:
> >>>> * fix some format issue
> >>>> --------
> >>>> ---
> >>>> sound/soc/codecs/rt715-sdca.c | 16 ++++++++++++++++
> >>>> sound/soc/codecs/rt715-sdca.h | 1 +
> >>>> sound/soc/codecs/rt715.c | 14 ++++++++++++++
> >>>> sound/soc/codecs/rt715.h | 1 +
> >>>> 4 files changed, 32 insertions(+)
> >>>>
> >>>> diff --git a/sound/soc/codecs/rt715-sdca.c b/sound/soc/codecs/rt715-sdca.c
> >>>> index b43ac8559e45..861a0d2a8957 100644
> >>>> --- a/sound/soc/codecs/rt715-sdca.c
> >>>> +++ b/sound/soc/codecs/rt715-sdca.c
> >>>> @@ -12,6 +12,7 @@
> >>>> #include <linux/version.h>
> >>>> #include <linux/kernel.h>
> >>>> #include <linux/init.h>
> >>>> +#include <linux/leds.h>
> >>>> #include <linux/pm_runtime.h>
> >>>> #include <linux/pm.h>
> >>>> #include <linux/soundwire/sdw.h>
> >>>> @@ -244,6 +245,7 @@ static int rt715_sdca_get_volsw(struct snd_kcontrol
> >>>> *kcontrol,
> >>>> unsigned int max = mc->max;
> >>>> int val;
> >>>>
> >>>> + pr_err("++++++rt715_sdca_get_volsw++\n");
> >>>> val = snd_soc_component_read(component, mc->reg);
> >>>> if (val < 0)
> >>>> return -EINVAL;
> >>>> @@ -261,6 +263,7 @@ static int rt715_sdca_put_volsw(struct snd_kcontrol
> >>>> *kcontrol,
> >>>> struct snd_ctl_elem_value *ucontrol)
> >>>> {
> >>>> struct snd_soc_component *component = snd_kcontrol_chip(kcontrol);
> >>>> + struct rt715_sdca_priv *rt715 =
> >>>> snd_soc_component_get_drvdata(component);
> >>>> struct soc_mixer_control *mc =
> >>>> (struct soc_mixer_control *)kcontrol->private_value;
> >>>> unsigned int val, val2, loop_cnt = 2, i;
> >>>> @@ -268,6 +271,7 @@ static int rt715_sdca_put_volsw(struct snd_kcontrol
> >>>> *kcontrol,
> >>>> unsigned int reg2 = mc->rreg;
> >>>> unsigned int reg = mc->reg;
> >>>> unsigned int max = mc->max;
> >>>> + unsigned int val0, val1;
> >>>> int err;
> >>>>
> >>>> val = ucontrol->value.integer.value[0];
> >>>> @@ -287,6 +291,18 @@ static int rt715_sdca_put_volsw(struct snd_kcontrol
> >>>> *kcontrol,
> >>>> return err;
> >>>> }
> >>>>
> >>>> +#if IS_ENABLED(CONFIG_DELL_PRIVACY)
> >>>> + /* dell privacy LED trigger state changed by muted/unmute switch */
> >>>> + if (mc->invert) {
> >>>> + if (ucontrol->value.integer.value[0] || ucontrol-
> >>>>> value.integer.value[1]) {
> >>>> + rt715->micmute_led = LED_OFF;
> >>>> + } else {
> >>>> + rt715->micmute_led = LED_ON;
> >>>> + }
> >>>> + ledtrig_audio_set(LED_AUDIO_MICMUTE, rt715->micmute_led);
> >>>> + }
> >>>> +#endif
> >>>> +
> >>>> return 0;
> >>>> }
> >>>>
> >>>> diff --git a/sound/soc/codecs/rt715-sdca.h b/sound/soc/codecs/rt715-sdca.h
> >>>> index 840c237895dd..f8988ab88f80 100644
> >>>> --- a/sound/soc/codecs/rt715-sdca.h
> >>>> +++ b/sound/soc/codecs/rt715-sdca.h
> >>>> @@ -31,6 +31,7 @@ struct rt715_sdca_priv {
> >>>> int l_is_unmute;
> >>>> int r_is_unmute;
> >>>> int hw_sdw_ver;
> >>>> + bool micmute_led;
> >>>> };
> >>>>
> >>>> struct rt715_sdw_stream_data {
> >>>> diff --git a/sound/soc/codecs/rt715.c b/sound/soc/codecs/rt715.c
> >>>> index cdcba70146da..b4e480744c94 100644
> >>>> --- a/sound/soc/codecs/rt715.c
> >>>> +++ b/sound/soc/codecs/rt715.c
> >>>> @@ -13,6 +13,7 @@
> >>>> #include <linux/init.h>
> >>>> #include <linux/delay.h>
> >>>> #include <linux/i2c.h>
> >>>> +#include <linux/leds.h>
> >>>> #include <linux/pm_runtime.h>
> >>>> #include <linux/pm.h>
> >>>> #include <linux/soundwire/sdw.h>
> >>>> @@ -88,6 +89,7 @@ static int rt715_set_amp_gain_put(struct snd_kcontrol
> >>>> *kcontrol,
> >>>> RT715_SET_GAIN_MIX_ADC2_L};
> >>>> unsigned int addr_h, addr_l, val_h, val_ll, val_lr;
> >>>> unsigned int read_ll, read_rl, i, j, loop_cnt;
> >>>> + unsigned int val0, val1;
> >>>>
> >>>> if (strstr(ucontrol->id.name, "Main Capture Switch") ||
> >>>> strstr(ucontrol->id.name, "Main Capture Volume"))
> >>>> @@ -95,6 +97,18 @@ static int rt715_set_amp_gain_put(struct snd_kcontrol
> >>>> *kcontrol,
> >>>> else
> >>>> loop_cnt = 1;
> >>>>
> >>>> +#if IS_ENABLED(CONFIG_DELL_PRIVACY)
> >>>> + /* Micmute LED state changed by muted/unmute switch */
> >>>> + if (mc->invert) {
> >>>> + if (ucontrol->value.integer.value[0] || ucontrol-
> >>>>> value.integer.value[1]) {
> >>>> + rt715->micmute_led = LED_OFF;
> >>>> + } else {
> >>>> + rt715->micmute_led = LED_ON;
> >>>> + }
> >>>> + ledtrig_audio_set(LED_AUDIO_MICMUTE, rt715->micmute_led);
> >>>> + }
> >>>> +#endif
> >>>> +
> >>> You might have missed my other comment on v2 feedback, but is there a reason
> >>> to keep it behind a compile time flag for dell privacy module? In practice
> >>> any other future led backend provider should work too. Another way to think
> >>> about it - if dell privacy wasn't enabled would this cause a problem to run
> >>> this code? I think it would just be a no-op.
> >>>
> >>>> for (j = 0; j < loop_cnt; j++) {
> >>>> /* Can't use update bit function, so read the original value first
> >>>> */
> >>>> if (loop_cnt == 1) {
> >>>> diff --git a/sound/soc/codecs/rt715.h b/sound/soc/codecs/rt715.h
> >>>> index 009a8266f606..57c9af041181 100644
> >>>> --- a/sound/soc/codecs/rt715.h
> >>>> +++ b/sound/soc/codecs/rt715.h
> >>>> @@ -22,6 +22,7 @@ struct rt715_priv {
> >>>> struct sdw_bus_params params;
> >>>> bool hw_init;
> >>>> bool first_hw_init;
> >>>> + bool micmute_led;
> >>>> };
> >>>>
> >>>> struct sdw_stream_data {
> >>>> --
> >>>> 2.25.1
> >>
> >> Pierre Louis suggested to just set the mic mute led state unconditionally .
> >> It is more common interface to allow other platforms to change micmute led.
> >> The discussion result is from "[PATCH v2 2/2] ASoC: rt715:add Mic Mute LED
> >> control support"
> >>
> >> Here is the change compared to V3 ,the CONFIG_DELL_PRIVACY will be removed
> >>
> >> -#if IS_ENABLED(CONFIG_DELL_PRIVACY)
> >> - /* dell privacy LED trigger state changed by muted/unmute switch */
> >> + /* MicMute LED state changed by muted/unmute switch */
> >> if (mc->invert) {
> >> if (ucontrol->value.integer.value[0] || ucontrol-
> >>> value.integer.value[1]) {
> >> rt715->micmute_led = LED_OFF;
> >> diff --git a/sound/soc/codecs/rt715.c b/sound/soc/codecs/rt715.c
> >> index b4e480744c94..60bb3d98103e 100644
> >> --- a/sound/soc/codecs/rt715.c
> >> +++ b/sound/soc/codecs/rt715.c
> >> @@ -97,8 +97,7 @@ static int rt715_set_amp_gain_put(struct snd_kcontrol
> >> *kcontrol,
> >> else
> >> loop_cnt = 1;
> >>
> >> -#if IS_ENABLED(CONFIG_DELL_PRIVACY)
> >> - /* Micmute LED state changed by muted/unmute switch */
> >> + /* MicMute LED state changed by muted/unmute switch */
> >> if (mc->invert) {
> >> if (ucontrol->value.integer.value[0] || ucontrol-
> >>> value.integer.value[1]) {
> >> rt715->micmute_led = LED_OFF;
> >> @@ -107,7 +106,6 @@ static int rt715_set_amp_gain_put(struct snd_kcontrol
> >> *kcontrol,
> >> }
> >> ledtrig_audio_set(LED_AUDIO_MICMUTE, rt715->micmute_led);
> >> }
> >> -#endif
> >
> > I'm not sure which thread, but I believe there was still also another suggestion
> > that this should be made more "generic" and to work for all codecs. So that when
> > we have lets say a hypothetical rt717 we don't need the same in that module.
>
> Right:
>
>
> On 1/15/21 6:45 PM, Jaroslav Kysela wrote:
> > Dne 12. 01. 21 v 18:18 Perry Yuan napsal(a):
> >
> >> +#if IS_ENABLED(CONFIG_DELL_PRIVACY)
> >> + /* dell privacy LED trigger state changed by muted/unmute switch */
> >> + if (mc->invert) {
> >> + if (ucontrol->value.integer.value[0] || ucontrol->value.integer.value[1]) {
> >> + rt715->micmute_led = LED_OFF;
> >> + } else {
> >> + rt715->micmute_led = LED_ON;
> >> + }
> >> + ledtrig_audio_set(LED_AUDIO_MICMUTE, rt715->micmute_led);
> >> + }
> >> +#endif
> >
> > [removing lkml and x86-platform lists]
> >
> > Looking to this code, perhaps, it may be a good idea to move the LED trigger
> > handling from the low-level drivers and do it in the top-level control code
> > (sound/core/control.c). Adding a new flag SNDRV_CTL_ELEM_ACCESS_MIC_LED may be
> > an option and we can nicely handle the situation for multiple microphones (OR
> > operation). The motivation is to remove the duplicated code.
> >
> > The same situation is for the playback LED.
>
> I believe that this is an excellent suggestion and that this is the way to go.
Well, I'm not thrilled by moving in that high level.
If you put it to the core, the control core (i.e. the main snd module)
will need the dependency on the LEDS stuff. Such an dependency mess
is what we'd like to avoid in general.
At most, yet more helper module that depends on the leds, but OTOH, we
don't want yet more module, either...
BTW, there is always some crazy request. For example, HD-audio code
has the mute-LED mode to flip the LED meaning; i.e. turn LED on at
unmuted so that user can see explicitly that it's being recorded.
It makes sense from the security POV.
Just to keep such a possibility in mind.
Takashi
