All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] ALSA: ASoC: cs4271: add optional soft reset workaround
@ 2012-12-10  9:30 Daniel Mack
  2012-12-10 20:58 ` Alexander Sverdlin
  2012-12-24 15:53 ` Mark Brown
  0 siblings, 2 replies; 3+ messages in thread
From: Daniel Mack @ 2012-12-10  9:30 UTC (permalink / raw)
  To: alsa-devel; +Cc: broonie, subaparts, lrg, Daniel Mack

The CS4271 requires its LRCLK and MCLK to be stable before its RESET
line is de-asserted. That also means that clocks cannot be changed
without putting the chip back into hardware reset, which also requires
a complete re-initialization of all registers.

One (undocumented) workaround is to assert and de-assert the PDN bit
in the MODE2 register.

This patch adds a new flag to both the DT bindings as well as to the
platform data to enable that workaround.

Signed-off-by: Daniel Mack <zonque@gmail.com>
---
 Documentation/devicetree/bindings/sound/cs4271.txt | 12 ++++++++
 include/sound/cs4271.h                             | 15 ++++++++++
 sound/soc/codecs/cs4271.c                          | 34 ++++++++++++++++++++++
 3 files changed, 61 insertions(+)

diff --git a/Documentation/devicetree/bindings/sound/cs4271.txt b/Documentation/devicetree/bindings/sound/cs4271.txt
index a850fb9..e2cd1d7 100644
--- a/Documentation/devicetree/bindings/sound/cs4271.txt
+++ b/Documentation/devicetree/bindings/sound/cs4271.txt
@@ -20,6 +20,18 @@ Optional properties:
 		!RESET pin
  - cirrus,amuteb-eq-bmutec:	When given, the Codec's AMUTEB=BMUTEC flag
 				is enabled.
+ - cirrus,enable-soft-reset:
+	The CS4271 requires its LRCLK and MCLK to be stable before its RESET
+	line is de-asserted. That also means that clocks cannot be changed
+	without putting the chip back into hardware reset, which also requires
+	a complete re-initialization of all registers.
+
+	One (undocumented) workaround is to assert and de-assert the PDN bit
+	in the MODE2 register. This workaround can be enabled with this DT
+	property.
+
+	Note that this is not needed in case the clocks are stable
+	throughout the entire runtime of the codec.
 
 Examples:
 
diff --git a/include/sound/cs4271.h b/include/sound/cs4271.h
index dd8c48d..70f4535 100644
--- a/include/sound/cs4271.h
+++ b/include/sound/cs4271.h
@@ -20,6 +20,21 @@
 struct cs4271_platform_data {
 	int gpio_nreset;	/* GPIO driving Reset pin, if any */
 	bool amutec_eq_bmutec;	/* flag to enable AMUTEC=BMUTEC */
+
+	/*
+	 * The CS4271 requires its LRCLK and MCLK to be stable before its RESET
+	 * line is de-asserted. That also means that clocks cannot be changed
+	 * without putting the chip back into hardware reset, which also requires
+	 * a complete re-initialization of all registers.
+	 *
+	 * One (undocumented) workaround is to assert and de-assert the PDN bit
+	 * in the MODE2 register. This workaround can be enabled with the
+	 * following flag.
+	 *
+	 * Note that this is not needed in case the clocks are stable
+	 * throughout the entire runtime of the codec.
+	 */
+	bool enable_soft_reset;
 };
 
 #endif /* __CS4271_H */
diff --git a/sound/soc/codecs/cs4271.c b/sound/soc/codecs/cs4271.c
index ac99238..ffa72dd 100644
--- a/sound/soc/codecs/cs4271.c
+++ b/sound/soc/codecs/cs4271.c
@@ -167,6 +167,8 @@ struct cs4271_private {
 	int				gpio_nreset;
 	/* GPIO that disable serial bus, if any */
 	int				gpio_disable;
+	/* enable soft reset workaround */
+	bool				enable_soft_reset;
 };
 
 /*
@@ -325,6 +327,33 @@ static int cs4271_hw_params(struct snd_pcm_substream *substream,
 	int i, ret;
 	unsigned int ratio, val;
 
+	if (cs4271->enable_soft_reset) {
+		/*
+		 * Put the codec in soft reset and back again in case it's not
+		 * currently streaming data. This way of bringing the codec in
+		 * sync to the current clocks is not explicitly documented in
+		 * the data sheet, but it seems to work fine, and in contrast
+		 * to a read hardware reset, we don't have to sync back all
+		 * registers every time.
+		 */
+
+		if ((substream->stream == SNDRV_PCM_STREAM_PLAYBACK &&
+		     !dai->capture_active) ||
+		    (substream->stream == SNDRV_PCM_STREAM_CAPTURE &&
+		     !dai->playback_active)) {
+			ret = snd_soc_update_bits(codec, CS4271_MODE2,
+						  CS4271_MODE2_PDN,
+						  CS4271_MODE2_PDN);
+			if (ret < 0)
+				return ret;
+
+			ret = snd_soc_update_bits(codec, CS4271_MODE2,
+						  CS4271_MODE2_PDN, 0);
+			if (ret < 0)
+				return ret;
+		}
+	}
+
 	cs4271->rate = params_rate(params);
 
 	/* Configure DAC */
@@ -484,6 +513,10 @@ static int cs4271_probe(struct snd_soc_codec *codec)
 		if (of_get_property(codec->dev->of_node,
 				     "cirrus,amutec-eq-bmutec", NULL))
 			amutec_eq_bmutec = true;
+
+		if (of_get_property(codec->dev->of_node,
+				     "cirrus,enable-soft-reset", NULL))
+			cs4271->enable_soft_reset = true;
 	}
 #endif
 
@@ -492,6 +525,7 @@ static int cs4271_probe(struct snd_soc_codec *codec)
 			gpio_nreset = cs4271plat->gpio_nreset;
 
 		amutec_eq_bmutec = cs4271plat->amutec_eq_bmutec;
+		cs4271->enable_soft_reset = cs4271plat->enable_soft_reset;
 	}
 
 	if (gpio_nreset >= 0)
-- 
1.7.11.7

^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] ALSA: ASoC: cs4271: add optional soft reset workaround
  2012-12-10  9:30 [PATCH] ALSA: ASoC: cs4271: add optional soft reset workaround Daniel Mack
@ 2012-12-10 20:58 ` Alexander Sverdlin
  2012-12-24 15:53 ` Mark Brown
  1 sibling, 0 replies; 3+ messages in thread
From: Alexander Sverdlin @ 2012-12-10 20:58 UTC (permalink / raw)
  To: Daniel Mack; +Cc: alsa-devel, broonie, lrg

Hi!

On Mon, 2012-12-10 at 10:30 +0100, Daniel Mack wrote:
> The CS4271 requires its LRCLK and MCLK to be stable before its RESET
> line is de-asserted. That also means that clocks cannot be changed
> without putting the chip back into hardware reset, which also requires
> a complete re-initialization of all registers.
> 
> One (undocumented) workaround is to assert and de-assert the PDN bit
> in the MODE2 register.
> 
> This patch adds a new flag to both the DT bindings as well as to the
> platform data to enable that workaround.
> 
> Signed-off-by: Daniel Mack <zonque@gmail.com>

Looks good to me.

Acked-by: Alexander Sverdlin <subaparts@yandex.ru>

--
Regards,
Alexander.

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] ALSA: ASoC: cs4271: add optional soft reset workaround
  2012-12-10  9:30 [PATCH] ALSA: ASoC: cs4271: add optional soft reset workaround Daniel Mack
  2012-12-10 20:58 ` Alexander Sverdlin
@ 2012-12-24 15:53 ` Mark Brown
  1 sibling, 0 replies; 3+ messages in thread
From: Mark Brown @ 2012-12-24 15:53 UTC (permalink / raw)
  To: Daniel Mack; +Cc: alsa-devel, subaparts, lrg


[-- Attachment #1.1: Type: text/plain, Size: 341 bytes --]

On Mon, Dec 10, 2012 at 10:30:04AM +0100, Daniel Mack wrote:
> The CS4271 requires its LRCLK and MCLK to be stable before its RESET
> line is de-asserted. That also means that clocks cannot be changed
> without putting the chip back into hardware reset, which also requires
> a complete re-initialization of all registers.

Applied, thanks.

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2012-12-24 15:53 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-12-10  9:30 [PATCH] ALSA: ASoC: cs4271: add optional soft reset workaround Daniel Mack
2012-12-10 20:58 ` Alexander Sverdlin
2012-12-24 15:53 ` Mark Brown

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.