* [PATCH] iio: light: bh1745: fix 16-bit register accesses
@ 2026-11-16 19:42 Arnav Kapoor
2026-11-17 10:21 ` Andy Shevchenko
0 siblings, 1 reply; 4+ messages in thread
From: Arnav Kapoor @ 2026-11-16 19:42 UTC (permalink / raw)
To: Jonathan Cameron
Cc: Mudit Sharma, David Lechner, Nuno Sá,
Andy Shevchenko, linux-iio, linux-kernel, Arnav Kapoor
The 16-bit colour data and threshold registers are little-endian
LSB/MSB pairs, but the driver accesses them through regmap_bulk_read()
and regmap_bulk_write() with an int or u16 buffer and no byte-order
conversion:
- bh1745_read_raw() reads 2 bytes into an uninitialised int, so the
upper 16 bits returned to user space are stack contents.
- bh1745_read_thresh() reads 2 bytes straight into *val, which the IIO
core leaves uninitialised, with the same result for the event
threshold attributes.
- bh1745_write_thresh() writes the first 2 bytes of an int, which are
the upper (zero) bytes on big-endian.
- bh1745_trigger_handler() stores the raw bytes in a u16 although the
scan type is IIO_CPU.
Use an __le16 buffer for all of them and convert with le16_to_cpu() /
cpu_to_le16().
Fixes: eab35358aae7 ("iio: light: ROHM BH1745 colour sensor")
Signed-off-by: Arnav Kapoor <kapoorarnav43@gmail.com>
---
Compile-tested only (W=1, checkpatch --strict clean); I don't have
BH1745 hardware. On little-endian the raw read currently returns the
register value plus whatever was on the stack in the upper 16 bits of
the uninitialised int; the threshold read has the same problem through
iio_ev_value_show(), whose val is not initialised either.
drivers/iio/light/bh1745.c | 29 +++++++++++++++++++----------
1 file changed, 19 insertions(+), 10 deletions(-)
diff --git a/drivers/iio/light/bh1745.c b/drivers/iio/light/bh1745.c
index 2b8ff36071b8..9c14499f848a 100644
--- a/drivers/iio/light/bh1745.c
+++ b/drivers/iio/light/bh1745.c
@@ -421,19 +421,20 @@ static int bh1745_read_raw(struct iio_dev *indio_dev,
int *val, int *val2, long mask)
{
struct bh1745_data *data = iio_priv(indio_dev);
+ __le16 value;
int ret;
- int value;
switch (mask) {
case IIO_CHAN_INFO_RAW:
if (!iio_device_claim_direct(indio_dev))
return -EBUSY;
- ret = regmap_bulk_read(data->regmap, chan->address, &value, 2);
+ ret = regmap_bulk_read(data->regmap, chan->address, &value,
+ sizeof(value));
iio_device_release_direct(indio_dev);
if (ret)
return ret;
- *val = value;
+ *val = le16_to_cpu(value);
return IIO_VAL_INT;
@@ -503,6 +504,7 @@ static int bh1745_read_thresh(struct iio_dev *indio_dev,
enum iio_event_info info, int *val, int *val2)
{
struct bh1745_data *data = iio_priv(indio_dev);
+ __le16 value;
int ret;
switch (info) {
@@ -510,18 +512,22 @@ static int bh1745_read_thresh(struct iio_dev *indio_dev,
switch (dir) {
case IIO_EV_DIR_RISING:
ret = regmap_bulk_read(data->regmap, BH1745_TH_LSB,
- val, 2);
+ &value, sizeof(value));
if (ret)
return ret;
+ *val = le16_to_cpu(value);
+
return IIO_VAL_INT;
case IIO_EV_DIR_FALLING:
ret = regmap_bulk_read(data->regmap, BH1745_TL_LSB,
- val, 2);
+ &value, sizeof(value));
if (ret)
return ret;
+ *val = le16_to_cpu(value);
+
return IIO_VAL_INT;
default:
@@ -547,6 +553,7 @@ static int bh1745_write_thresh(struct iio_dev *indio_dev,
enum iio_event_info info, int val, int val2)
{
struct bh1745_data *data = iio_priv(indio_dev);
+ __le16 value;
int ret;
switch (info) {
@@ -554,10 +561,12 @@ static int bh1745_write_thresh(struct iio_dev *indio_dev,
if (val < 0x0 || val > 0xFFFF)
return -EINVAL;
+ value = cpu_to_le16(val);
+
switch (dir) {
case IIO_EV_DIR_RISING:
ret = regmap_bulk_write(data->regmap, BH1745_TH_LSB,
- &val, 2);
+ &value, sizeof(value));
if (ret)
return ret;
@@ -565,7 +574,7 @@ static int bh1745_write_thresh(struct iio_dev *indio_dev,
case IIO_EV_DIR_FALLING:
ret = regmap_bulk_write(data->regmap, BH1745_TL_LSB,
- &val, 2);
+ &value, sizeof(value));
if (ret)
return ret;
@@ -741,18 +750,18 @@ static irqreturn_t bh1745_trigger_handler(int interrupt, void *p)
u16 chans[4];
aligned_s64 timestamp;
} scan = { };
- u16 value;
+ __le16 value;
int ret;
int i;
int j = 0;
iio_for_each_active_channel(indio_dev, i) {
ret = regmap_bulk_read(data->regmap, BH1745_RED_LSB + 2 * i,
- &value, 2);
+ &value, sizeof(value));
if (ret)
goto err;
- scan.chans[j++] = value;
+ scan.chans[j++] = le16_to_cpu(value);
}
iio_push_to_buffers_with_ts(indio_dev, &scan, sizeof(scan),
base-commit: a6d9114f68c4b5507b2d945bd6d1e37415f913c5
--
2.53.0
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] iio: light: bh1745: fix 16-bit register accesses
2026-11-16 19:42 [PATCH] iio: light: bh1745: fix 16-bit register accesses Arnav Kapoor
@ 2026-11-17 10:21 ` Andy Shevchenko
2026-11-17 10:41 ` Arnav Kapoor
0 siblings, 1 reply; 4+ messages in thread
From: Andy Shevchenko @ 2026-11-17 10:21 UTC (permalink / raw)
To: Arnav Kapoor
Cc: Jonathan Cameron, Mudit Sharma, David Lechner, Nuno Sá,
Andy Shevchenko, linux-iio, linux-kernel
On Tue, Nov 17, 2026 at 01:12:10AM +0530, Arnav Kapoor wrote:
> The 16-bit colour data and threshold registers are little-endian
> LSB/MSB pairs, but the driver accesses them through regmap_bulk_read()
> and regmap_bulk_write() with an int or u16 buffer and no byte-order
> conversion:
>
> - bh1745_read_raw() reads 2 bytes into an uninitialised int, so the
> upper 16 bits returned to user space are stack contents.
> - bh1745_read_thresh() reads 2 bytes straight into *val, which the IIO
> core leaves uninitialised, with the same result for the event
> threshold attributes.
> - bh1745_write_thresh() writes the first 2 bytes of an int, which are
> the upper (zero) bytes on big-endian.
> - bh1745_trigger_handler() stores the raw bytes in a u16 although the
> scan type is IIO_CPU.
I didn't get this, what's the problem here?
> Use an __le16 buffer for all of them and convert with le16_to_cpu() /
> cpu_to_le16().
Otherwise seems legit to me.
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
And you also need to add
asm/byteorder.h.
--
With Best Regards,
Andy Shevchenko
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] iio: light: bh1745: fix 16-bit register accesses
2026-11-17 10:21 ` Andy Shevchenko
@ 2026-11-17 10:41 ` Arnav Kapoor
2026-11-17 10:51 ` Andy Shevchenko
0 siblings, 1 reply; 4+ messages in thread
From: Arnav Kapoor @ 2026-11-17 10:41 UTC (permalink / raw)
To: Andy Shevchenko
Cc: Jonathan Cameron, Mudit Sharma, David Lechner, Nuno Sá,
Andy Shevchenko, linux-iio, linux-kernel
On Tue, Nov 17, 2026 at 12:21:42PM +0200, Andy Shevchenko wrote:
> On Tue, Nov 17, 2026 at 01:12:10AM +0530, Arnav Kapoor wrote:
> > - bh1745_trigger_handler() stores the raw bytes in a u16 although the
> > scan type is IIO_CPU.
>
> I didn't get this, what's the problem here?
The handler does regmap_bulk_read(..., &value, 2) into a u16 and puts
that straight into scan.chans[], so the buffer gets the register bytes
as they come off the bus, LSB first. The channels declare
.endianness = IIO_CPU, so user space reads each sample as a native u16.
That is right on little-endian, but on big-endian every sample comes
out byte-swapped. The patch reads into an __le16 and stores
le16_to_cpu(value), so the data matches IIO_CPU on both.
I'll reword that bullet in v2 to say this directly.
> Otherwise seems legit to me.
> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Thanks!
> And you also need to add
>
> asm/byteorder.h.
Will do in v2.
Thanks,
Arnav
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] iio: light: bh1745: fix 16-bit register accesses
2026-11-17 10:41 ` Arnav Kapoor
@ 2026-11-17 10:51 ` Andy Shevchenko
0 siblings, 0 replies; 4+ messages in thread
From: Andy Shevchenko @ 2026-11-17 10:51 UTC (permalink / raw)
To: Arnav Kapoor
Cc: Jonathan Cameron, Mudit Sharma, David Lechner, Nuno Sá,
Andy Shevchenko, linux-iio, linux-kernel
On Tue, Nov 17, 2026 at 04:11:52PM +0530, Arnav Kapoor wrote:
> On Tue, Nov 17, 2026 at 12:21:42PM +0200, Andy Shevchenko wrote:
> > On Tue, Nov 17, 2026 at 01:12:10AM +0530, Arnav Kapoor wrote:
...
> > > - bh1745_trigger_handler() stores the raw bytes in a u16 although the
> > > scan type is IIO_CPU.
> >
> > I didn't get this, what's the problem here?
> The handler does regmap_bulk_read(..., &value, 2) into a u16 and puts
> that straight into scan.chans[], so the buffer gets the register bytes
> as they come off the bus, LSB first. The channels declare
> .endianness = IIO_CPU, so user space reads each sample as a native u16.
> That is right on little-endian, but on big-endian every sample comes
> out byte-swapped. The patch reads into an __le16 and stores
> le16_to_cpu(value), so the data matches IIO_CPU on both.
>
> I'll reword that bullet in v2 to say this directly.
Yes, please do.
--
With Best Regards,
Andy Shevchenko
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2026-11-17 10:51 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-16 19:42 [PATCH] iio: light: bh1745: fix 16-bit register accesses Arnav Kapoor
2026-11-17 10:21 ` Andy Shevchenko
2026-11-17 10:41 ` Arnav Kapoor
2026-11-17 10:51 ` Andy Shevchenko
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®