[PATCH 1/2] ASoC: tlv320adcx140: Avoid accessing invalid gpio_reset
From: Camel Guo camelg@axis.com
When gpio_reset is not well defined in devicetree, the adcx140->gpio_reset is an error code instead of NULL. In this case, adcx140->gpio_reset should not be used by adcx140_reset. This commit sets it NULL to avoid accessing an invalid variable.
Signed-off-by: Camel Guo camelg@axis.com --- sound/soc/codecs/tlv320adcx140.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/sound/soc/codecs/tlv320adcx140.c b/sound/soc/codecs/tlv320adcx140.c index 7ae6ec374be3..597dd1062943 100644 --- a/sound/soc/codecs/tlv320adcx140.c +++ b/sound/soc/codecs/tlv320adcx140.c @@ -984,8 +984,10 @@ static int adcx140_i2c_probe(struct i2c_client *i2c,
adcx140->gpio_reset = devm_gpiod_get_optional(adcx140->dev, "reset", GPIOD_OUT_LOW); - if (IS_ERR(adcx140->gpio_reset)) + if (IS_ERR(adcx140->gpio_reset) || adcx140->gpio_reset == NULL) { dev_info(&i2c->dev, "Reset GPIO not defined\n"); + adcx140->gpio_reset = NULL; + }
adcx140->supply_areg = devm_regulator_get_optional(adcx140->dev, "areg");
From: Camel Guo camelg@axis.com
According to its datasheet, after reset this codec goes into sleep mode. In this mode, any register accessing should be avoided except for exiting sleep mode. Hence this commit moves SLEEP_CFG access before any register accessing.
Signed-off-by: Camel Guo camelg@axis.com --- sound/soc/codecs/tlv320adcx140.c | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/sound/soc/codecs/tlv320adcx140.c b/sound/soc/codecs/tlv320adcx140.c index 597dd1062943..6d456aa269ad 100644 --- a/sound/soc/codecs/tlv320adcx140.c +++ b/sound/soc/codecs/tlv320adcx140.c @@ -842,6 +842,18 @@ static int adcx140_codec_probe(struct snd_soc_component *component) if (ret) goto out;
+ if (adcx140->supply_areg == NULL) + sleep_cfg_val |= ADCX140_AREG_INTERNAL; + + ret = regmap_write(adcx140->regmap, ADCX140_SLEEP_CFG, sleep_cfg_val); + if (ret) { + dev_err(adcx140->dev, "setting sleep config failed %d\n", ret); + goto out; + } + + /* 8.4.3: Wait >= 1ms after entering active mode. */ + usleep_range(1000, 100000); + pdm_count = device_property_count_u32(adcx140->dev, "ti,pdm-edge-select"); if (pdm_count <= ADCX140_NUM_PDM_EDGES && pdm_count > 0) { @@ -889,18 +901,6 @@ static int adcx140_codec_probe(struct snd_soc_component *component) if (ret) goto out;
- if (adcx140->supply_areg == NULL) - sleep_cfg_val |= ADCX140_AREG_INTERNAL; - - ret = regmap_write(adcx140->regmap, ADCX140_SLEEP_CFG, sleep_cfg_val); - if (ret) { - dev_err(adcx140->dev, "setting sleep config failed %d\n", ret); - goto out; - } - - /* 8.4.3: Wait >= 1ms after entering active mode. */ - usleep_range(1000, 100000); - ret = regmap_update_bits(adcx140->regmap, ADCX140_BIAS_CFG, ADCX140_MIC_BIAS_VAL_MSK | ADCX140_MIC_BIAS_VREF_MSK, bias_cfg);
Camel
On 9/8/20 3:35 AM, Camel Guo wrote:
From: Camel Guo camelg@axis.com
According to its datasheet, after reset this codec goes into sleep mode. In this mode, any register accessing should be avoided except for exiting sleep mode. Hence this commit moves SLEEP_CFG access before any register accessing.
This is interesting because our HW team suggested putting the device into sleep mode when doing register writes/reads because they were finding abnormalities in the register settings when the device is active.
I have a local patch that changes this as well that the HW team requested.
Mark
Let me run this by the HW team first before applying this patch.
Dan
Camel
On 9/8/20 6:49 AM, Dan Murphy wrote:
Camel
On 9/8/20 3:35 AM, Camel Guo wrote:
From: Camel Guo camelg@axis.com
According to its datasheet, after reset this codec goes into sleep mode. In this mode, any register accessing should be avoided except for exiting sleep mode. Hence this commit moves SLEEP_CFG access before any register accessing.
This is interesting because our HW team suggested putting the device into sleep mode when doing register writes/reads because they were finding abnormalities in the register settings when the device is active.
I have a local patch that changes this as well that the HW team requested.
OK I have clarification on this now. Their original request was incorrect they indicate the BIAS, ADC and PLLs be powered down during writes and reads.
Mark
Let me run this by the HW team first before applying this patch.
Mark
Acked-by: Dan Murphy dmurphy@ti.com
Dan
Camel
On 9/8/20 3:35 AM, Camel Guo wrote:
From: Camel Guo camelg@axis.com
When gpio_reset is not well defined in devicetree, the adcx140->gpio_reset is an error code instead of NULL. In this case, adcx140->gpio_reset should not be used by adcx140_reset. This commit sets it NULL to avoid accessing an invalid variable.
Signed-off-by: Camel Guo camelg@axis.com
sound/soc/codecs/tlv320adcx140.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/sound/soc/codecs/tlv320adcx140.c b/sound/soc/codecs/tlv320adcx140.c index 7ae6ec374be3..597dd1062943 100644 --- a/sound/soc/codecs/tlv320adcx140.c +++ b/sound/soc/codecs/tlv320adcx140.c @@ -984,8 +984,10 @@ static int adcx140_i2c_probe(struct i2c_client *i2c,
adcx140->gpio_reset = devm_gpiod_get_optional(adcx140->dev, "reset", GPIOD_OUT_LOW);
- if (IS_ERR(adcx140->gpio_reset))
- if (IS_ERR(adcx140->gpio_reset) || adcx140->gpio_reset == NULL) {
This looks a bit off and seems like the NULL check was added just to print the message.
I would suggest removing the "or" check and just set the gpio_reset to NULL in an error case.
This avoids noise in the log especially if the gpio_reset is intentionally not populated in the DT
Dan
On 08/09/2020 11.35, Camel Guo wrote:
From: Camel Guo camelg@axis.com
When gpio_reset is not well defined in devicetree, the adcx140->gpio_reset is an error code instead of NULL. In this case, adcx140->gpio_reset should not be used by adcx140_reset. This commit sets it NULL to avoid accessing an invalid variable.
Signed-off-by: Camel Guo camelg@axis.com
sound/soc/codecs/tlv320adcx140.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/sound/soc/codecs/tlv320adcx140.c b/sound/soc/codecs/tlv320adcx140.c index 7ae6ec374be3..597dd1062943 100644 --- a/sound/soc/codecs/tlv320adcx140.c +++ b/sound/soc/codecs/tlv320adcx140.c @@ -984,8 +984,10 @@ static int adcx140_i2c_probe(struct i2c_client *i2c,
adcx140->gpio_reset = devm_gpiod_get_optional(adcx140->dev, "reset", GPIOD_OUT_LOW);
- if (IS_ERR(adcx140->gpio_reset))
- if (IS_ERR(adcx140->gpio_reset) || adcx140->gpio_reset == NULL) { dev_info(&i2c->dev, "Reset GPIO not defined\n");
adcx140->gpio_reset = NULL;
the correct fix is to: dev_err(&i2c->dev, "Reset GPIO not defined\n"); return PTR_ERR(adcx140->gpio_reset);
If the reset GPIO is specified and you get error when requesting it as optional, there is a reason for that. For example deferred probing.
}
adcx140->supply_areg = devm_regulator_get_optional(adcx140->dev, "areg");
- Péter
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
Hi,
On 08/09/2020 14.59, Peter Ujfalusi wrote:
On 08/09/2020 11.35, Camel Guo wrote:
From: Camel Guo camelg@axis.com
When gpio_reset is not well defined in devicetree, the adcx140->gpio_reset is an error code instead of NULL. In this case, adcx140->gpio_reset should not be used by adcx140_reset. This commit sets it NULL to avoid accessing an invalid variable.
Signed-off-by: Camel Guo camelg@axis.com
sound/soc/codecs/tlv320adcx140.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/sound/soc/codecs/tlv320adcx140.c b/sound/soc/codecs/tlv320adcx140.c index 7ae6ec374be3..597dd1062943 100644 --- a/sound/soc/codecs/tlv320adcx140.c +++ b/sound/soc/codecs/tlv320adcx140.c @@ -984,8 +984,10 @@ static int adcx140_i2c_probe(struct i2c_client *i2c,
adcx140->gpio_reset = devm_gpiod_get_optional(adcx140->dev, "reset", GPIOD_OUT_LOW);
- if (IS_ERR(adcx140->gpio_reset))
- if (IS_ERR(adcx140->gpio_reset) || adcx140->gpio_reset == NULL) { dev_info(&i2c->dev, "Reset GPIO not defined\n");
adcx140->gpio_reset = NULL;the correct fix is to: dev_err(&i2c->dev, "Reset GPIO not defined\n");
no need to print, I think gpio core will do that.
return PTR_ERR(adcx140->gpio_reset);
If the reset GPIO is specified and you get error when requesting it as optional, there is a reason for that. For example deferred probing.
}
adcx140->supply_areg = devm_regulator_get_optional(adcx140->dev, "areg");
- Péter
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
- Péter
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
On Tue, 8 Sep 2020 10:35:20 +0200, Camel Guo wrote:
When gpio_reset is not well defined in devicetree, the adcx140->gpio_reset is an error code instead of NULL. In this case, adcx140->gpio_reset should not be used by adcx140_reset. This commit sets it NULL to avoid accessing an invalid variable.
Applied to
https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next
Thanks!
[1/1] ASoC: tlv320adcx140: Wake up codec before accessing register commit: 1a5ce48fd667128e369fdc7fb87e21539aed21b5
All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.
You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.
If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.
Please add any relevant lists and maintainers to the CCs when replying to this mail.
Thanks, Mark
participants (4)
-
Camel Guo -
Dan Murphy -
Mark Brown -
Peter Ujfalusi