[PATCH v2] ASoC: fsl_sai: Fix one bug for hardware limitation.

[PATCH v2] ASoC: fsl_sai: Fix one bug for hardware limitation.

[Xiubo Li]

This is maybe one bug or a limitation of the hardware that the {T,R}CR2's
Synchronous Mode bits must be set as late as possible, or the SAI device
maybe hanged up, and there has not any explaination about this limitation
in the SAI Data Sheet.

And the {T,R}CR2's Synchronous Mode bits must be set at the same time whether
for Tx or Rx stream.

Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
---
 sound/soc/fsl/fsl_sai.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/sound/soc/fsl/fsl_sai.c b/sound/soc/fsl/fsl_sai.c
index af80246..2ece147 100644
--- a/sound/soc/fsl/fsl_sai.c
+++ b/sound/soc/fsl/fsl_sai.c
@@ -145,7 +145,6 @@ static int fsl_sai_set_dai_fmt_tr(struct snd_soc_dai *cpu_dai,
 	switch (fmt & SND_SOC_DAIFMT_FORMAT_MASK) {
 	case SND_SOC_DAIFMT_I2S:
 		val_cr4 |= FSL_SAI_CR4_FSE;
-		val_cr4 |= FSL_SAI_CR4_FSP;
 		break;
 	default:
 		return -EINVAL;
@@ -185,9 +184,6 @@ static int fsl_sai_set_dai_fmt_tr(struct snd_soc_dai *cpu_dai,
 		return -EINVAL;
 	}
 
-	if (fsl_dir == FSL_FMT_RECEIVER)
-		val_cr2 |= FSL_SAI_CR2_SYNC;
-
 	sai_writel(sai, val_cr2, sai->base + reg_cr2);
 	sai_writel(sai, val_cr4, sai->base + reg_cr4);
 
@@ -253,6 +249,7 @@ static int fsl_sai_hw_params(struct snd_pcm_substream *substream,
 	val_cr5 |= FSL_SAI_CR5_WNW(word_width);
 	val_cr5 |= FSL_SAI_CR5_W0W(word_width);
 
+	val_cr5 &= ~FSL_SAI_CR5_FBT_MASK;
 	if (sai->big_endian_data)
 		val_cr5 |= FSL_SAI_CR5_FBT(word_width - 1);
 	else
@@ -272,7 +269,15 @@ static int fsl_sai_trigger(struct snd_pcm_substream *substream, int cmd,
 		struct snd_soc_dai *cpu_dai)
 {
 	struct fsl_sai *sai = snd_soc_dai_get_drvdata(cpu_dai);
-	u32 tcsr, rcsr, val_cr3, reg_cr3;
+	u32 tcsr, rcsr, val_cr2, val_cr3, reg_cr3;
+
+	val_cr2 = sai_readl(sai, sai->base + FSL_SAI_TCR2);
+	val_cr2 &= ~FSL_SAI_CR2_SYNC;
+	sai_writel(sai, val_cr2, sai->base + FSL_SAI_TCR2);
+
+	val_cr2 = sai_readl(sai, sai->base + FSL_SAI_RCR2);
+	val_cr2 |= FSL_SAI_CR2_SYNC;
+	sai_writel(sai, val_cr2, sai->base + FSL_SAI_RCR2);
 
 	tcsr = sai_readl(sai, sai->base + FSL_SAI_TCSR);
 	rcsr = sai_readl(sai, sai->base + FSL_SAI_RCSR);
-- 
1.8.4

[PATCH] ASoC: fsl_sai: fix the endianess for SAI fifo data.

[Xiubo Li]

Revert the SAI's endianess for fifo data to/from DMA engine.

Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
---
 sound/soc/fsl/fsl_sai.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/sound/soc/fsl/fsl_sai.c b/sound/soc/fsl/fsl_sai.c
index 2ece147..5d38a67 100644
--- a/sound/soc/fsl/fsl_sai.c
+++ b/sound/soc/fsl/fsl_sai.c
@@ -138,9 +138,9 @@ static int fsl_sai_set_dai_fmt_tr(struct snd_soc_dai *cpu_dai,
 	val_cr4 = sai_readl(sai, sai->base + reg_cr4);
 
 	if (sai->big_endian_data)
-		val_cr4 |= FSL_SAI_CR4_MF;
-	else
 		val_cr4 &= ~FSL_SAI_CR4_MF;
+	else
+		val_cr4 |= FSL_SAI_CR4_MF;
 
 	switch (fmt & SND_SOC_DAIFMT_FORMAT_MASK) {
 	case SND_SOC_DAIFMT_I2S:
@@ -251,9 +251,9 @@ static int fsl_sai_hw_params(struct snd_pcm_substream *substream,
 
 	val_cr5 &= ~FSL_SAI_CR5_FBT_MASK;
 	if (sai->big_endian_data)
-		val_cr5 |= FSL_SAI_CR5_FBT(word_width - 1);
-	else
 		val_cr5 |= FSL_SAI_CR5_FBT(0);
+	else
+		val_cr5 |= FSL_SAI_CR5_FBT(word_width - 1);
 
 	val_cr4 |= FSL_SAI_CR4_FRSZ(channels);
 	val_mr = ~0UL - ((1 << channels) - 1);
-- 
1.8.4

Re: [PATCH v2] ASoC: fsl_sai: Fix one bug for hardware limitation.

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 328 bytes --]

On Tue, Dec 31, 2013 at 03:33:21PM +0800, Xiubo Li wrote:
> This is maybe one bug or a limitation of the hardware that the {T,R}CR2's
> Synchronous Mode bits must be set as late as possible, or the SAI device
> maybe hanged up, and there has not any explaination about this limitation
> in the SAI Data Sheet.

Applied, thanks.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] ASoC: fsl_sai: fix the endianess for SAI fifo data.

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 139 bytes --]

On Tue, Dec 31, 2013 at 03:33:22PM +0800, Xiubo Li wrote:
> Revert the SAI's endianess for fifo data to/from DMA engine.

Applied, thanks.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]