Skip to contentSkip to conversation

ALSA: hda/tas2781: Fix sound abnormal issue on some SPI device

PATCH In review Baojun Xu 3 participants 4 messages Apr 18 – Apr 20

Subsystems (4)
lore
Share & tools
lore

Baojun Xu

author maintainer
#0 lore raw

PATCH 1/1 ALSA: hda/tas2781: Fix sound abnormal issue on some SPI device

In the SPI driver probe, the chip ID must be set to TAS2781. Without this initialization, calibration data fails to load correctly, causing audio abnormalities on some devices.
And update the register bulk read API to handle the distinct requirements of SPI and I2C devices.

Fixes 05ac3846ffe5 ("ALSA: hda/tas2781: A workaround solution to lower-vol issue among lower calibrated-impedance micro-speaker on TAS2781")
---
 sound/hda/codecs/side-codecs/tas2781_hda_spi.c | 1 +
 sound/soc/codecs/tas2781-fmwlib.c              | 7 +++----
 2 files changed, 4 insertions(+), 4 deletions(-)
sound/hda/codecs/side-codecs/tas2781_hda_spi.c +1 -0
index f860e0eb7602..560f2385212d 100644
@@ -788,6 +788,7 @@ static int tas2781_hda_spi_probe(struct spi_device *spi)
 	}
 	if (strstr(dev_name(&spi->dev), "TXNW2781")) {
 		device_name = "TXNW2781";
+		tas_hda->priv->chip_id = TAS2781;
 	} else {
 		dev_err(tas_priv->dev, "Unmatched spi dev %s\n",
 			dev_name(&spi->dev));
sound/soc/codecs/tas2781-fmwlib.c +3 -4
index a1d86bd309f4..885e0b6fed00 100644
@@ -2487,7 +2487,7 @@ static int tas2781_cali_preproc(struct tasdevice_priv *priv, int i)
 		if (spec == NULL)
 			return -ENOMEM;
 		priv->tasdevice[i].cali_specific = spec;
-		rc = tasdevice_dev_bulk_read(priv, i, p->r0_reg, r0_deflt, 4);
+		rc = priv->dev_bulk_read(priv, i, p->r0_reg, r0_deflt, 4);
 		if (rc < 0) {
 			dev_err(priv->dev, "invalid RE from %d = %d\n", i, rc);
 			return rc;
@@ -2511,9 +2511,8 @@ static int tas2781_cali_preproc(struct tasdevice_priv *priv, int i)
 				TASDEVICE_REG(0, 0x1b, 0x34) :
 				TASDEVICE_REG(0, 0x18, 0x1c);
 
-			rc = tasdevice_dev_bulk_read(priv, i,
-						      spec->sin_gni_reg,
-						      spec->sin_gni, 4);
+			rc = priv->dev_bulk_read(priv, i, spec->sin_gni_reg,
+						 spec->sin_gni, 4);
 			if (rc < 0) {
 				dev_err(priv->dev, "wrong sinegaini %d = %d\n",
 					i, rc);
Signature · 1 line
2.25.1
#1 lore raw
Quoted 9 lines · Baojun Xu

In the SPI driver probe, the chip ID must be set to TAS2781. Without this initialization, calibration data fails to load correctly, causing audio abnormalities on some devices.
And update the register bulk read API to handle the distinct requirements of SPI and I2C devices.

Fixes 05ac3846ffe5 ("ALSA: hda/tas2781: A workaround solution to lower-vol issue among lower calibrated-impedance micro-speaker on TAS2781")

Applied now. Thanks.

Takashi

#2 lore raw
In reply to Baojun Xu
Quoted 40 lines · Baojun Xu

In the SPI driver probe, the chip ID must be set to TAS2781. Without this initialization, calibration data fails to load correctly, causing audio abnormalities on some devices.
And update the register bulk read API to handle the distinct requirements of SPI and I2C devices.

Fixes 05ac3846ffe5 ("ALSA: hda/tas2781: A workaround solution to lower-vol issue among lower calibrated-impedance micro-speaker on TAS2781")
---
 sound/hda/codecs/side-codecs/tas2781_hda_spi.c | 1 +
 sound/soc/codecs/tas2781-fmwlib.c              | 7 +++----
 2 files changed, 4 insertions(+), 4 deletions(-)
sound/hda/codecs/side-codecs/tas2781_hda_spi.c +1 -0
index f860e0eb7602..560f2385212d 100644
@@ -788,6 +788,7 @@ static int tas2781_hda_spi_probe(struct spi_device *spi)
 	}
 	if (strstr(dev_name(&spi->dev), "TXNW2781")) {
 		device_name = "TXNW2781";
+		tas_hda->priv->chip_id = TAS2781;
 	} else {
 		dev_err(tas_priv->dev, "Unmatched spi dev %s\n",
 			dev_name(&spi->dev));
sound/soc/codecs/tas2781-fmwlib.c +1 -1
index a1d86bd309f4..885e0b6fed00 100644
@@ -2487,7 +2487,7 @@ static int tas2781_cali_preproc(struct tasdevice_priv *priv, int i)
 		if (spec == NULL)
 			return -ENOMEM;
 		priv->tasdevice[i].cali_specific = spec;
-		rc = tasdevice_dev_bulk_read(priv, i, p->r0_reg, r0_deflt, 4);
+		rc = priv->dev_bulk_read(priv, i, p->r0_reg, r0_deflt, 4);
 		if (rc < 0) {
 			dev_err(priv->dev, "invalid RE from %d = %d\n", i, rc);
 			return rc;
@@ -2511,9 +2511,8 @@ static int tas2781_cali_preproc(struct tasdevice_priv *priv, int i)
 				TASDEVICE_REG(0, 0x1b, 0x34) :

sound/soc/codecs/tas2781-fmwlib.c static int tas2781_cali_preproc(struct tasdevice_priv *priv, int i) · line 2518

Quoted 10 lines
sound/soc/codecs/tas2781-fmwlib.c +2 -3
 				TASDEVICE_REG(0, 0x18, 0x1c);
 
-			rc = tasdevice_dev_bulk_read(priv, i,
-						      spec->sin_gni_reg,
-						      spec->sin_gni, 4);
+			rc = priv->dev_bulk_read(priv, i, spec->sin_gni_reg,
+						 spec->sin_gni, 4);
 			if (rc < 0) {
 				dev_err(priv->dev, "wrong sinegaini %d = %d\n",
 					i, rc);

Hi Baojun, will this fix
https://lore.kernel.org/linux-sound/[email protected]/ ?

Similarly, I noticed that something like the above was merged upstream, commit 0bdf27abaf894.

Signature · 2 lines
Regards,
  Eric
author maintainer
#3 lore raw

Hi Eric,

Quoted 15 lines

________________________________________
From: Eric Naim <[email protected]>
Sent: 20 April 2026 14:43
To: Xu, Baojun; [email protected]; [email protected]
Cc: [email protected]; [email protected]; [email protected]; Ding, Shenghao; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: [EXTERNAL] Re: [PATCH v1] ALSA: hda/tas2781: Fix sound abnormal issue on some SPI device

Quoted 6 lines · Baojun Xu

In the SPI driver probe, the chip ID must be set to TAS2781. Without this initialization, calibration data fails to load correctly, causing audio abnormalities on some devices.
And update the register bulk read API to handle the distinct requirements of SPI and I2C devices.

Quoted 9 lines
Quoted 3 lines
Fixes 05ac3846ffe5 ("ALSA: hda/tas2781: A workaround solution to lower-vol issue among lower calibrated-impedance micro-speaker on TAS2781")

---

Hi Baojun, will this fix

https://urldefense.com/v3/__https://lore.kernel.org/linux-sound/[email protected]/__;!!G3vK!X-jfopjNDyUV4Zt7eHbLDT79zxX8PrVATeyHGzNtgCNvue6wSwqHWSaaVOuqjRfYbzZ5TQzNVhRq$ ?

Similarly, I noticed that something like the above was merged upstream, commit 0bdf27abaf894.

To clarify: The patch mentioned earlier was designed to register the correct project with the codec. My patch specifically resolves the issue where calibration data fails to be applied properly.

Quoted 5 lines

--
Regards,
Eric

Best Regards
Jim