[PATCH 0/3] iio: pressure: dps310: fix input checks and resets

[PATCH 0/3] iio: pressure: dps310: fix input checks and resets

[Rupesh Majhi]

Sashiko flagged these as pre-existing in its review of the FIFO series
v8 [1]. All three reproduce on a DPS310 on a BeagleBone Black:

- 0 or 0.5 written to a rate or ratio is accepted. Ratio 0 puts a
  reserved value in the register: pressure reads fail with -ERANGE,
  temperature reads 974 C.
- A warm reboot during a FIFO capture leaves FIFO_EN set. After it,
  every raw read times out until the driver is unbound.
- The reset after a read timeout silently puts the rate and ratio back
  to their defaults.

Patch 2 fixes a commit that is only in iio testing so far.

Paulius, patch 2 also resets the ZTS8032 at probe, the same reset that
remove already does. A Tested-by would be welcome if you have time.

Based on iio testing. Reproduced and fixed with help from an LLM coding
assistant.

[1] https://sashiko.dev/#/patchset/20260921183132.233136-1-zoone.rupert%40gmail.com

Cc: Paulius Zaleckas <paulius.zaleckas@gmail.com>

Rupesh Majhi (3):
  iio: pressure: dps310: reject a zero rate or oversampling ratio
  iio: pressure: dps310: reset the chip at probe
  iio: pressure: dps310: keep the user's settings across a timeout reset

 drivers/iio/pressure/dps310.c | 38 ++++++++++++++++++++++++++---------
 1 file changed, 28 insertions(+), 10 deletions(-)


base-commit: 7d80b4b383bc53b48abd3a30ca4a91cdab568bfa
-- 
2.43.0

[PATCH 1/3] iio: pressure: dps310: reject a zero rate or oversampling ratio

[Rupesh Majhi]

The rate and ratio setters accept 0, which also covers 0.5. ilog2(0)
gives -1, which masked into the register is 128 Hz or the reserved
precision 0b1111. The latter makes every pressure read fail with
-ERANGE and temperature read 974 degrees C.

Reject values below 1.

Fixes: ba6ec48e76bc ("iio: Add driver for Infineon DPS310")
Fixes: d711a3c7dc82 ("iio: dps310: Add pressure sensing capability")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260921183132.233136-1-zoone.rupert%40gmail.com
Signed-off-by: Rupesh Majhi <zoone.rupert@gmail.com>
---
Cc: Paulius Zaleckas <paulius.zaleckas@gmail.com>

 drivers/iio/pressure/dps310.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/iio/pressure/dps310.c b/drivers/iio/pressure/dps310.c
index 71053f4463e5..b6aa5dbe2bcd 100644
--- a/drivers/iio/pressure/dps310.c
+++ b/drivers/iio/pressure/dps310.c
@@ -410,7 +410,7 @@ static int dps310_set_pres_precision(struct dps310_data *data, int val)
 	int rc;
 	u8 shift_en;
 
-	if (val < 0 || val > 128)
+	if (val < 1 || val > 128)
 		return -EINVAL;
 
 	shift_en = val >= 16 ? DPS310_PRS_SHIFT_EN : 0;
@@ -433,7 +433,7 @@ static int dps310_set_temp_precision(struct dps310_data *data, int val)
 	int rc;
 	u8 shift_en;
 
-	if (val < 0 || val > 128)
+	if (val < 1 || val > 128)
 		return -EINVAL;
 
 	shift_en = val >= 16 ? DPS310_TMP_SHIFT_EN : 0;
@@ -455,7 +455,7 @@ static int dps310_set_pres_samp_freq(struct dps310_data *data, int freq)
 {
 	u8 val;
 
-	if (freq < 0 || freq > 128)
+	if (freq < 1 || freq > 128)
 		return -EINVAL;
 
 	val = ilog2(freq) << 4;
@@ -469,7 +469,7 @@ static int dps310_set_temp_samp_freq(struct dps310_data *data, int freq)
 {
 	u8 val;
 
-	if (freq < 0 || freq > 128)
+	if (freq < 1 || freq > 128)
 		return -EINVAL;
 
 	val = ilog2(freq) << 4;
-- 
2.43.0

[PATCH 2/3] iio: pressure: dps310: reset the chip at probe

[Rupesh Majhi]

The chip keeps its registers across a warm reboot, and probe changes
only the shift bits in CFG_REG. A reboot during a FIFO capture leaves
FIFO_EN set, so the ready bits never come and every raw read times
out. The timeout reset does not clear it either, since the CFG_REG
update that follows writes FIFO_EN back from the register cache.

Reset the chip first, as remove already does.

Fixes: 90f0d1016398 ("iio: pressure: dps310: read buffered samples from the hardware FIFO")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260921183132.233136-1-zoone.rupert%40gmail.com
Signed-off-by: Rupesh Majhi <zoone.rupert@gmail.com>
---
Cc: Paulius Zaleckas <paulius.zaleckas@gmail.com>

 drivers/iio/pressure/dps310.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/iio/pressure/dps310.c b/drivers/iio/pressure/dps310.c
index b6aa5dbe2bcd..29778f25bb33 100644
--- a/drivers/iio/pressure/dps310.c
+++ b/drivers/iio/pressure/dps310.c
@@ -1445,6 +1445,11 @@ static int dps310_probe(struct i2c_client *client)
 	if (rc)
 		return rc;
 
+	/* The chip keeps its state across a warm reboot */
+	rc = dps310_reset_wait(data);
+	if (rc)
+		return rc;
+
 	rc = dps310_startup(data);
 	if (rc)
 		return rc;
-- 
2.43.0

[PATCH 3/3] iio: pressure: dps310: keep the user's settings across a timeout reset

[Rupesh Majhi]

When a sample never becomes ready, the driver resets the chip and runs
startup again, which writes the default rate and ratio. With 64x at
4 Hz set and the chip put in standby, the next read recovers, but both
channels are silently back to 1x at 1 Hz.

The register cache still holds the user's values. Write them back
after the reset, before measuring starts again, and move the defaults
out of dps310_startup() so only probe writes them.

A rate and ratio the chip cannot keep up with, such as 128x at 128 Hz,
now keep timing out instead of falling back to 1x at 1 Hz.

Fixes: 7b4ab4abcea4 ("iio: pressure: dps310: Reset chip after timeout")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260921183132.233136-1-zoone.rupert%40gmail.com
Signed-off-by: Rupesh Majhi <zoone.rupert@gmail.com>
---
Cc: Paulius Zaleckas <paulius.zaleckas@gmail.com>

 drivers/iio/pressure/dps310.c | 25 +++++++++++++++++++------
 1 file changed, 19 insertions(+), 6 deletions(-)

diff --git a/drivers/iio/pressure/dps310.c b/drivers/iio/pressure/dps310.c
index 29778f25bb33..c364fbaeb3fe 100644
--- a/drivers/iio/pressure/dps310.c
+++ b/drivers/iio/pressure/dps310.c
@@ -306,10 +306,9 @@ static int dps310_temp_workaround(struct dps310_data *data)
 	return regmap_write(data->regmap, 0x0f, 0x00);
 }
 
-static int dps310_startup(struct dps310_data *data)
+static int dps310_set_defaults(struct dps310_data *data)
 {
 	int rc;
-	int ready;
 
 	/*
 	 * Set up pressure sensor in single sample, one measurement per second
@@ -328,10 +327,14 @@ static int dps310_startup(struct dps310_data *data)
 		return rc;
 
 	/* Temp and pressure shifts are disabled when PRC <= 8 */
-	rc = regmap_write_bits(data->regmap, DPS310_CFG_REG,
-			       DPS310_PRS_SHIFT_EN | DPS310_TMP_SHIFT_EN, 0);
-	if (rc)
-		return rc;
+	return regmap_write_bits(data->regmap, DPS310_CFG_REG,
+				 DPS310_PRS_SHIFT_EN | DPS310_TMP_SHIFT_EN, 0);
+}
+
+static int dps310_startup(struct dps310_data *data)
+{
+	int rc;
+	int ready;
 
 	/* MEAS_CFG doesn't update correctly unless first written with 0 */
 	rc = regmap_write_bits(data->regmap, DPS310_MEAS_CFG,
@@ -551,6 +554,12 @@ static int dps310_reset_reinit(struct dps310_data *data)
 	if (rc)
 		return rc;
 
+	/* Put back the configuration the reset cleared */
+	regcache_mark_dirty(data->regmap);
+	rc = regcache_sync_region(data->regmap, DPS310_PRS_CFG, DPS310_CFG_REG);
+	if (rc)
+		return rc;
+
 	return dps310_startup(data);
 }
 
@@ -1450,6 +1459,10 @@ static int dps310_probe(struct i2c_client *client)
 	if (rc)
 		return rc;
 
+	rc = dps310_set_defaults(data);
+	if (rc)
+		return rc;
+
 	rc = dps310_startup(data);
 	if (rc)
 		return rc;
-- 
2.43.0

Re: [PATCH 1/3] iio: pressure: dps310: reject a zero rate or oversampling ratio

[Joshua Crofts]

On Fri, 30 Oct 2026 00:06:33 +0200
Rupesh Majhi <zoone.rupert@gmail.com> wrote:

> The rate and ratio setters accept 0, which also covers 0.5. ilog2(0)
> gives -1, which masked into the register is 128 Hz or the reserved
> precision 0b1111. The latter makes every pressure read fail with
> -ERANGE and temperature read 974 degrees C.
> 
> Reject values below 1.
> 
> Fixes: ba6ec48e76bc ("iio: Add driver for Infineon DPS310")
> Fixes: d711a3c7dc82 ("iio: dps310: Add pressure sensing capability")
> Reported-by: Sashiko <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260921183132.233136-1-zoone.rupert%40gmail.com
> Signed-off-by: Rupesh Majhi <zoone.rupert@gmail.com>
> ---

Reviewed-by: Joshua Crofts <crofts@kernel.org>

-- 
Kind regards,
Joshua Crofts

Re: [PATCH 2/3] iio: pressure: dps310: reset the chip at probe

[Joshua Crofts]

On Fri, 30 Oct 2026 00:06:34 +0200
Rupesh Majhi <zoone.rupert@gmail.com> wrote:

> The chip keeps its registers across a warm reboot, and probe changes
> only the shift bits in CFG_REG. A reboot during a FIFO capture leaves
> FIFO_EN set, so the ready bits never come and every raw read times
> out. The timeout reset does not clear it either, since the CFG_REG
> update that follows writes FIFO_EN back from the register cache.
> 
> Reset the chip first, as remove already does.
> 
> Fixes: 90f0d1016398 ("iio: pressure: dps310: read buffered samples from the hardware FIFO")
> Reported-by: Sashiko <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260921183132.233136-1-zoone.rupert%40gmail.com
> Signed-off-by: Rupesh Majhi <zoone.rupert@gmail.com>
> ---
> Cc: Paulius Zaleckas <paulius.zaleckas@gmail.com>
> 
>  drivers/iio/pressure/dps310.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/iio/pressure/dps310.c b/drivers/iio/pressure/dps310.c
> index b6aa5dbe2bcd..29778f25bb33 100644
> --- a/drivers/iio/pressure/dps310.c
> +++ b/drivers/iio/pressure/dps310.c
> @@ -1445,6 +1445,11 @@ static int dps310_probe(struct i2c_client *client)
>  	if (rc)
>  		return rc;
>  
> +	/* The chip keeps its state across a warm reboot */
> +	rc = dps310_reset_wait(data);
> +	if (rc)
> +		return rc;
> +
>  	rc = dps310_startup(data);
>  	if (rc)
>  		return rc;

Reviewed-by: Joshua Crofts <crofts@kernel.org>

-- 
Kind regards,
Joshua Crofts

Re: [PATCH 3/3] iio: pressure: dps310: keep the user's settings across a timeout reset

[Joshua Crofts]

On Fri, 30 Oct 2026 00:06:35 +0200
Rupesh Majhi <zoone.rupert@gmail.com> wrote:

...

> @@ -328,10 +327,14 @@ static int dps310_startup(struct dps310_data *data)
>  		return rc;
>  
>  	/* Temp and pressure shifts are disabled when PRC <= 8 */
> -	rc = regmap_write_bits(data->regmap, DPS310_CFG_REG,
> -			       DPS310_PRS_SHIFT_EN | DPS310_TMP_SHIFT_EN, 0);
> -	if (rc)
> -		return rc;
> +	return regmap_write_bits(data->regmap, DPS310_CFG_REG,
> +				 DPS310_PRS_SHIFT_EN | DPS310_TMP_SHIFT_EN, 0);
> +}
> +
> +static int dps310_startup(struct dps310_data *data)
> +{
> +	int rc;
> +	int ready;

A very small nit, but reverse xmas tree.

-- 
Kind regards,
Joshua Crofts

Re: [PATCH 2/3] iio: pressure: dps310: reset the chip at probe

[Jonathan Cameron]

On Fri, 30 Oct 2026 00:06:34 +0200
Rupesh Majhi <zoone.rupert@gmail.com> wrote:

> The chip keeps its registers across a warm reboot, and probe changes
> only the shift bits in CFG_REG. A reboot during a FIFO capture leaves
> FIFO_EN set, so the ready bits never come and every raw read times
> out. The timeout reset does not clear it either, since the CFG_REG
> update that follows writes FIFO_EN back from the register cache.
> 
> Reset the chip first, as remove already does.
> 
> Fixes: 90f0d1016398 ("iio: pressure: dps310: read buffered samples from the hardware FIFO")
That's not a stable hash because it is recent enough that
it is on the tree I've rebased since then and will do again on
rc1.  Given this will go in at same time as the fifo anyway
during next merge cycle, I tend to just drop the fixes tags
in cases like this to avoid having to fix them up.

Other than that, series looks fine to me, but given you asked Paulius
for a test I'll wait on that!

Thanks

Jonathan


> Reported-by: Sashiko <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260921183132.233136-1-zoone.rupert%40gmail.com
> Signed-off-by: Rupesh Majhi <zoone.rupert@gmail.com>
> ---
> Cc: Paulius Zaleckas <paulius.zaleckas@gmail.com>
> 
>  drivers/iio/pressure/dps310.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/iio/pressure/dps310.c b/drivers/iio/pressure/dps310.c
> index b6aa5dbe2bcd..29778f25bb33 100644
> --- a/drivers/iio/pressure/dps310.c
> +++ b/drivers/iio/pressure/dps310.c
> @@ -1445,6 +1445,11 @@ static int dps310_probe(struct i2c_client *client)
>  	if (rc)
>  		return rc;
>  
> +	/* The chip keeps its state across a warm reboot */
> +	rc = dps310_reset_wait(data);
> +	if (rc)
> +		return rc;
> +
>  	rc = dps310_startup(data);
>  	if (rc)
>  		return rc;

Re: [PATCH 2/3] iio: pressure: dps310: reset the chip at probe

[Jonathan Cameron]

On Fri, 30 Oct 2026 00:06:34 +0200
Rupesh Majhi <zoone.rupert@gmail.com> wrote:

> The chip keeps its registers across a warm reboot, and probe changes
> only the shift bits in CFG_REG. A reboot during a FIFO capture leaves
> FIFO_EN set, so the ready bits never come and every raw read times
> out. The timeout reset does not clear it either, since the CFG_REG
> update that follows writes FIFO_EN back from the register cache.
> 
> Reset the chip first, as remove already does.
Why do we do it in remove?  Does it put it in a power saving mode
or similar? It's fairly unusual to bother resetting in that path
given as you've seen here, we pretty much have to reset in probe.

Jonathan

> 
> Fixes: 90f0d1016398 ("iio: pressure: dps310: read buffered samples from the hardware FIFO")
> Reported-by: Sashiko <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260921183132.233136-1-zoone.rupert%40gmail.com
> Signed-off-by: Rupesh Majhi <zoone.rupert@gmail.com>
> ---
> Cc: Paulius Zaleckas <paulius.zaleckas@gmail.com>
> 
>  drivers/iio/pressure/dps310.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/iio/pressure/dps310.c b/drivers/iio/pressure/dps310.c
> index b6aa5dbe2bcd..29778f25bb33 100644
> --- a/drivers/iio/pressure/dps310.c
> +++ b/drivers/iio/pressure/dps310.c
> @@ -1445,6 +1445,11 @@ static int dps310_probe(struct i2c_client *client)
>  	if (rc)
>  		return rc;
>  
> +	/* The chip keeps its state across a warm reboot */
> +	rc = dps310_reset_wait(data);
> +	if (rc)
> +		return rc;
> +
>  	rc = dps310_startup(data);
>  	if (rc)
>  		return rc;