This series adds support for the MAX40080, a bidirectional current-sense
amplifier with an integrated 12-bit ADC.

The driver is placed in IIO rather than hwmon because:
  - IIO's scale/raw model naturally exposes the selectable gain ranges and
    lets userspace apply the correct scale per range.
  - Buffered capture support is planned for a follow-up series -- each
    conversion returns a matched current+voltage pair, which maps cleanly
    to IIO's buffered/triggered model.

Changes since v7:
  - Add array_size.h include (Andy).
  - Add asm/byteorder.h include, use __le32 buf with le32_to_cpu()
    instead of u8 buf[4] with get_unaligned_le32() (Andy).
  - Use BIT_ULL() instead of 1ULL * BIT() (Andy).
  - Use READ_ONCE/WRITE_ONCE for st->range access (Jonathan).
  - Move propname declaration to function top, assignment near use (Andy).
  - Remove unnecessary blank lines between variable assignments and
    their use (Andy).
  - Place dev_err_probe on single line (Andy).
  - Fix Kconfig/Makefile alphabetical ordering -- MAX40080 between
    MAX34408 and MAX77541 (Andy).

Changes since v6:
  - Add types.h include (Andy).
  - Remove dead code: i2c_smbus_read_i2c_block_data can't return partial
    reads, so remove the ret != sizeof(buf) check (David).
  - Remove dead code: iv variable initialization not needed as
    max40080_read_iv always writes to *iv before returning (David).

Changes since v5:
  - Fix datasheet URL format in binding -- use stable link without /en/
    path component (Jonathan, Siratul).
  - Use 'adi' compatible prefix instead of 'maxim' -- MAX40080 is an ADI
    product (Conor).
  - Split MAINTAINERS into binding and driver patches (Siratul).
  - Add datasheet link in driver header comment (Siratul).
  - Fix comment style: capital first letter + period (Siratul).
  - Fix register defines: add register name prefix (CFG_, FIFO_CFG_, IV_)
    and use designated initializers for csa_gain array (Jonathan).
  - Define MAX40080_ADC_RES_BITS and use BIT() for shift (Jonathan).
  - Use lowercase 'mV' in defines (Siratul).
  - Use C99 loop variable declarations (Siratul).
  - Remove braces from single-statement if blocks (Siratul).
  - Declaration and initialization in same statement (Siratul).
  - Use IIO_VAL_FRACTIONAL_LOG2 for voltage scale (Jonathan).
  - Remove unused i2c_set_clientdata call (Jonathan).
  - Use defines instead of magic numbers for range default (Jonathan).

Changes since v4:
  - Initialize tmp = 0 in max40080_read_iv() before read_poll_timeout to
    avoid using potentially uninitialized value in the condition check
    (Sashiko, Jonathan).
  - Simplify max40080_get_oversampling_ratio() to use cached value instead
    of reading from the device (Jonathan).
  - Remove DMA-safe buffer from struct -- I2C layer already bounces data
    internally for i2c_smbus_* calls (Jonathan).
  - Make shunt-resistor-micro-ohms required in DT binding. The default
    1 Ohm was arbitrary and rarely correct for real hardware (Jonathan).
  - Remove PM ops entirely. Having only resume without suspend was flagged
    as "papering over" possible power loss issues (Jonathan).
  - Remove the {} scope around propname, declare it at function top (Andy).
  - Remove mod_devicetable.h include -- i2c.h provides necessary types
    (Jonathan).

Changes since v3:
  - Removed all regmap usage -- regmap doesn't support SMBus PEC or Quick
    Command, both required by this device (Nuno).
  - Fixed locking: max40080_update_bits() now takes the lock internally,
    and callers update cached state after the call returns (Nuno).
  - Consistent comment style: capital first letter + period (Andy).
  - Use propname variable to deduplicate "shunt-resistor-micro-ohms" (Andy).
  - Return IIO_VAL_* inline instead of break + return pattern (Nuno).

Changes since v2:
  - Added shunt-resistor-micro-ohms DT property with default value.
  - Use single-shot mode with SMBus Quick Command trigger for correlated
    current/voltage reads.
  - Precompute current scale per range at probe time.

Link: https://lore.kernel.org/linux-iio/20260818142928.8244-1-stefan.popa@analog.com/ [v7]
Link: https://lore.kernel.org/linux-iio/20260730131738.822405-1-stefan.popa@analog.com/ [v6]
Link: https://lore.kernel.org/linux-iio/20260723065036.2683075-1-stefan.popa@analog.com/ [v5]
Link: https://lore.kernel.org/linux-iio/20260717123852.1140893-1-stefan.popa@analog.com/ [v4]
Link: https://lore.kernel.org/linux-iio/20260715063652.368501-1-stefan.popa@analog.com/ [v3]
Link: https://lore.kernel.org/linux-iio/20260713120226.90303-1-stefan.popa@analog.com/ [v2]
Link: https://lore.kernel.org/linux-iio/20260703102941.1141341-1-stefan.popa@analog.com/ [v1]

Stefan Popa (2):
  dt-bindings: iio: adc: add adi,max40080
  iio: adc: add MAX40080 current-sense amplifier driver

 .../bindings/iio/adc/adi,max40080.yaml        |  65 ++
 MAINTAINERS                                   |   9 +
 drivers/iio/adc/Kconfig                       |  11 +
 drivers/iio/adc/Makefile                      |   1 +
 drivers/iio/adc/max40080.c                    | 572 ++++++++++++++++++
 5 files changed, 658 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/iio/adc/adi,max40080.yaml
 create mode 100644 drivers/iio/adc/max40080.c

--
2.53.0

> This series adds support for the MAX40080, a bidirectional current-sense
> amplifier with an integrated 12-bit ADC.
> 
> The driver is placed in IIO rather than hwmon because:
>   - IIO's scale/raw model naturally exposes the selectable gain ranges and
>     lets userspace apply the correct scale per range.

This one is a bit weaker as autoscaling is possible (if fiddly).

>   - Buffered capture support is planned for a follow-up series -- each
>     conversion returns a matched current+voltage pair, which maps cleanly
>     to IIO's buffered/triggered model.

If you have this ready to post, I'd prefer seeing it sooner rather than
later just to strengthen the arguement.  A link to a repo would do as well.

Anyhow, just a nice to have!

-- 
Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>

On Sun, 30 Aug 2026 01:51:39 +0100
Jonathan Cameron <jic23@kernel.org> wrote:

> > This series adds support for the MAX40080, a bidirectional current-sense
> > amplifier with an integrated 12-bit ADC.
> > 
> > The driver is placed in IIO rather than hwmon because:
> >   - IIO's scale/raw model naturally exposes the selectable gain ranges and
> >     lets userspace apply the correct scale per range.  
> 
> This one is a bit weaker as autoscaling is possible (if fiddly).
> 
> >   - Buffered capture support is planned for a follow-up series -- each
> >     conversion returns a matched current+voltage pair, which maps cleanly
> >     to IIO's buffered/triggered model.  
> 
> If you have this ready to post, I'd prefer seeing it sooner rather than
> later just to strengthen the arguement.  A link to a repo would do as well.
> 
> Anyhow, just a nice to have!
> 

The set of changes that Andy and I have asked for are pretty minor
so I would just have made them as tweaks whilst applying.  However, others
(e.g. David) have been active in reviewing this series in early versions
so I want to leave it on list anyway for some more time.

Given Andy's request wasn't entirely trivial, if you have time in a few
days to do a v9 with that stuff tidied up that'll save me a few mins
and cut down the possibility of me messing things up!

Jonathan