Thread (4 messages) 4 messages, 3 authors, 2025-10-04

Re: [PATCH v2] iio: adc: mt6360: Handle error in cleanup path correctly

From: Jonathan Cameron <jic23@kernel.org>
Date: 2025-10-04 14:31:30
Also in: linux-iio, linux-mediatek, lkml

On Tue, 30 Sep 2025 00:24:53 +0800
Haotian Zhang [off-list ref] wrote:
The return value of a regmap_raw_write() and regmap_update_bits()
in the cleanup path was being ignored.

Fix this by checking the return value and warn on error.

Fixes: 1f4877218f7e ("iio: adc: mt6360: Add ADC driver for MT6360")
Signed-off-by: Haotian Zhang <redacted>

---
Changes in v2:
 - Do not propagate cleanup path errors.
 - Log a warning on failure instead of overwriting the return value, as
   suggested by the maintainer.
As below. I think dev_err() is appropriate.
quoted hunk ↗ jump to hunk
 - Also check the return value of regmap_update_bits() for consistency.
---
 drivers/iio/adc/mt6360-adc.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers/iio/adc/mt6360-adc.c b/drivers/iio/adc/mt6360-adc.c
index 69b3569c90e5..9ee7247aacbe 100644
--- a/drivers/iio/adc/mt6360-adc.c
+++ b/drivers/iio/adc/mt6360-adc.c
@@ -70,6 +70,7 @@ static int mt6360_adc_read_channel(struct mt6360_adc_data *mad, int channel, int
 	ktime_t predict_end_t, timeout;
 	unsigned int pre_wait_time;
 	int ret;
+	int cleanup_ret;
 
 	mutex_lock(&mad->adc_lock);
 
@@ -130,11 +131,16 @@ static int mt6360_adc_read_channel(struct mt6360_adc_data *mad, int channel, int
 out_adc_conv:
 	/* Only keep ADC enable */
 	adc_enable = cpu_to_be16(MT6360_ADCEN_MASK);
-	regmap_raw_write(mad->regmap, MT6360_REG_PMUADCCFG, &adc_enable, sizeof(adc_enable));
+	cleanup_ret = regmap_raw_write(mad->regmap, MT6360_REG_PMUADCCFG,
+				&adc_enable, sizeof(adc_enable));
+	if (cleanup_ret)
+		dev_warn(mad->dev, "Failed to reset ADC config: %d\n", cleanup_ret);
Why warn? If this happens it's definite and error and may bite us later.
 	mad->last_off_timestamps[channel] = ktime_get();
 	/* Config prefer channel to NO_PREFER */
-	regmap_update_bits(mad->regmap, MT6360_REG_PMUADCRPT1, MT6360_PREFERCH_MASK,
+	cleanup_ret = regmap_update_bits(mad->regmap, MT6360_REG_PMUADCRPT1, MT6360_PREFERCH_MASK,
 			   MT6360_NO_PREFER << MT6360_PREFERCH_SHFT);
+	if (cleanup_ret)
+		dev_warn(mad->dev, "Failed to reset prefer channel: %d\n", cleanup_ret);
 out_adc_lock:
 	mutex_unlock(&mad->adc_lock);
 
Maybe it is worth trying to surface the error if nothing else has already gone wrong. That is a little fiddly
to do but something like

	if (cleanup_ret) {
		dev_err()
		ret = ret ?: cleanup_ret;
	}

I'm not sure it is worth the complexity however, so perhaps see if others have comments on this
in next few days before spinning a v3.

Thanks,

Jonathan
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help