[PATCH 1/2] ASoC: remove dapm_update_bits()

[PATCH 1/2] ASoC: remove dapm_update_bits()

[Lu Guanqun]

dapm_update_bits() is only called in dapm_generic_apply_power().  And in turn,
dapm_generic_apply_power() is only called in dapm_seq_run() when w->id is
snd_soc_dapm_{input|output|hp|line|mic}. Under these cases, dapm_update_bits()
is rather useless, because the below tests are always true:
        /* check for valid widgets */
        if (widget->reg < 0 || widget->id == snd_soc_dapm_input ||
                widget->id == snd_soc_dapm_output ||
                widget->id == snd_soc_dapm_hp ||
                widget->id == snd_soc_dapm_mic ||
                widget->id == snd_soc_dapm_line ||
                widget->id == snd_soc_dapm_spk)
                return 0;
Therefore, we better omit this function for simplicity.

Signed-off-by: Lu Guanqun <guanqun.lu@intel.com>
---
 sound/soc/soc-dapm.c |   41 -----------------------------------------
 1 files changed, 0 insertions(+), 41 deletions(-)

diff --git a/sound/soc/soc-dapm.c b/sound/soc/soc-dapm.c
index 81c4052..d511bbe 100644
--- a/sound/soc/soc-dapm.c
+++ b/sound/soc/soc-dapm.c
@@ -322,45 +322,6 @@ static int dapm_connect_mixer(struct snd_soc_dapm_context *dapm,
 	return -ENODEV;
 }
 
-/* update dapm codec register bits */
-static int dapm_update_bits(struct snd_soc_dapm_widget *widget)
-{
-	int change, power;
-	unsigned int old, new;
-	struct snd_soc_codec *codec = widget->codec;
-	struct snd_soc_dapm_context *dapm = widget->dapm;
-	struct snd_soc_card *card = dapm->card;
-
-	/* check for valid widgets */
-	if (widget->reg < 0 || widget->id == snd_soc_dapm_input ||
-		widget->id == snd_soc_dapm_output ||
-		widget->id == snd_soc_dapm_hp ||
-		widget->id == snd_soc_dapm_mic ||
-		widget->id == snd_soc_dapm_line ||
-		widget->id == snd_soc_dapm_spk)
-		return 0;
-
-	power = widget->power;
-	if (widget->invert)
-		power = (power ? 0:1);
-
-	old = snd_soc_read(codec, widget->reg);
-	new = (old & ~(0x1 << widget->shift)) | (power << widget->shift);
-
-	change = old != new;
-	if (change) {
-		pop_dbg(dapm->dev, card->pop_time,
-			"pop test %s : %s in %d ms\n",
-			widget->name, widget->power ? "on" : "off",
-			card->pop_time);
-		pop_wait(card->pop_time);
-		snd_soc_write(codec, widget->reg, new);
-	}
-	dev_dbg(dapm->dev, "reg %x old %x new %x change %d\n", widget->reg,
-		old, new, change);
-	return change;
-}
-
 /* create new dapm mixer control */
 static int dapm_new_mixer(struct snd_soc_dapm_context *dapm,
 	struct snd_soc_dapm_widget *w)
@@ -673,8 +634,6 @@ static int dapm_generic_apply_power(struct snd_soc_dapm_widget *w)
 			return ret;
 	}
 
-	dapm_update_bits(w);
-
 	/* power up post event */
 	if (w->power && w->event &&
 	    (w->event_flags & SND_SOC_DAPM_POST_PMU)) {
-- 
1.7.2.3

[PATCH 2/2] ASoC: rename dapm_generic_apply_power()

[Lu Guanqun]

dapm_generic_apply_power() is only called in dapm_seq_run() when w->id is
snd_soc_dapm_{input|output|hp|line|mic}. So, it's not a generic function at
all, we better rename it to reflect what it's doing.

Signed-off-by: Lu Guanqun <guanqun.lu@intel.com>
---
 sound/soc/soc-dapm.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/sound/soc/soc-dapm.c b/sound/soc/soc-dapm.c
index d511bbe..9b7ac3f 100644
--- a/sound/soc/soc-dapm.c
+++ b/sound/soc/soc-dapm.c
@@ -605,10 +605,10 @@ int dapm_reg_event(struct snd_soc_dapm_widget *w,
 }
 EXPORT_SYMBOL_GPL(dapm_reg_event);
 
-/* Standard power change method, used to apply power changes to most
- * widgets.
+/*
+ * power change method, used for input, output, hp, line, spk, mic.
  */
-static int dapm_generic_apply_power(struct snd_soc_dapm_widget *w)
+static int dapm_input_output_apply_power(struct snd_soc_dapm_widget *w)
 {
 	int ret;
 
@@ -947,7 +947,7 @@ static void dapm_seq_run(struct snd_soc_dapm_context *dapm,
 		case snd_soc_dapm_line:
 		case snd_soc_dapm_spk:
 			/* No register support currently */
-			ret = dapm_generic_apply_power(w);
+			ret = dapm_input_output_apply_power(w);
 			break;
 
 		default:
-- 
1.7.2.3

Re: [PATCH 2/2] ASoC: rename dapm_generic_apply_power()

[Mark Brown]

On Sat, Apr 02, 2011 at 03:23:36PM +0800, Lu Guanqun wrote:
> dapm_generic_apply_power() is only called in dapm_seq_run() when w->id is
> snd_soc_dapm_{input|output|hp|line|mic}. So, it's not a generic function at
> all, we better rename it to reflect what it's doing.

I don't think you've really understood what the code is doing here.
While it's true that we don't ever call this function for most widget
types due to the write compression code it is actually the generic
method for applying power changes to a widget in isolation - it does a
register update surrounded by the event callbacks in the appropriate
sequence which really is something that is generic to all widgets.  A
better question if you want to refector the code here would be why these
widgets are treated differently to other widgets.

Re: [PATCH 2/2] ASoC: rename dapm_generic_apply_power()

[Lu Guanqun]

Hi Mark,

On Sat, Apr 02, 2011 at 03:35:29PM +0800, Mark Brown wrote:
> On Sat, Apr 02, 2011 at 03:23:36PM +0800, Lu Guanqun wrote:
> > dapm_generic_apply_power() is only called in dapm_seq_run() when w->id is
> > snd_soc_dapm_{input|output|hp|line|mic}. So, it's not a generic function at
> > all, we better rename it to reflect what it's doing.
> 
> I don't think you've really understood what the code is doing here.
> While it's true that we don't ever call this function for most widget
> types due to the write compression code it is actually the generic

I don't get into write compression code yet, do you mean by "coalesce
register writes", that part of code?

> method for applying power changes to a widget in isolation - it does a
> register update surrounded by the event callbacks in the appropriate
> sequence which really is something that is generic to all widgets.  A
> better question if you want to refector the code here would be why these
> widgets are treated differently to other widgets.

-- 
guanqun

Re: [PATCH 2/2] ASoC: rename dapm_generic_apply_power()

[Mark Brown]

On Sat, Apr 02, 2011 at 03:44:33PM +0800, Lu Guanqun wrote:
> On Sat, Apr 02, 2011 at 03:35:29PM +0800, Mark Brown wrote:

> > I don't think you've really understood what the code is doing here.
> > While it's true that we don't ever call this function for most widget
> > types due to the write compression code it is actually the generic

> I don't get into write compression code yet, do you mean by "coalesce
> register writes", that part of code?

Yes.