AD5313R device and triggered buffer support

[rodrigo455]

PR Description

This series adds support for the AD5313R 10-bit nanoDAC and triggered
buffer support to the ad5686 DAC driver family, along with a number of
driver cleanups.

Initial patches update the device-tree bindings:

• Add AD5313R compatible entry (SPI variant)
• Add missing AD5673R/AD5677R entries (I2C variant)
• Add LDAC GPIO property for async DAC channel loading

Driver cleanups:

• Refactor include headers (IWYU)
• Switch to devm_mutex_init()
• Drop enum chip id in favor of per-device chip_info structs
  exposed to the SPI/I2C bus drivers
• Introduce bus ops struct with a sync() operation for batching
  bus transfers

New functionality:

• AD5313R device support
• LDAC GPIO handling (active-low, held low when unused)
• SPI bus sync() implementation for batching multiple transfers
• Triggered buffer support, leveraging LDAC and sync() to flush
  all channel writes atomically

PR Type

• Bug fix (a change that fixes an issue)
• New feature (a change that adds new functionality)
• Breaking change (a change that affects other repos or cause CIs to fail)

PR Checklist

• I have conducted a self-review of my own code changes
• I have compiled my changes, including the documentation
• I have tested the changes on the relevant hardware
• I have updated the documentation outside this repo accordingly
• I have provided links for the relevant upstream lore

[nunojsa]

Looks mostly good! Some things more relevant to change but I would say it's mostly upstream ready.

Thanks for taking care of this and bringing one more driver up to date!

[nunojsa]

nit: new line

I also wonder if this does not need locking?

[github-actions]

Automated Kernel Review — AD5313R IIO DAC Series

CI run: https://github.com/analogdevicesinc/linux/actions/runs/24441066096

Reviewed commits aa19856315a7..e05048f77b21 (14 code patches + 1 b4 cover letter).

Build passes cleanly for ARM (W=1, no warnings). DT binding checks (dt_binding_check) pass with no errors.

---

🐛 Bug: bus_data->size not reset on spi_sync() failure

drivers/iio/dac/ad5686-spi.c, commit 5a685dae

// Current (broken on error path):
ret = spi_sync(spi, &bus_data->msg);
if (ret)
    return ret;          // bus_data->size is NOT reset

bus_data->size = 0;
return 0;

If spi_sync() fails, bus_data->size is left non-zero. On the next call to ad5686_spi_write(), the stale size > 0 check causes spi_message_init() to be skipped, and new transfers are appended to the already-executed (and failed) message — producing a corrupt SPI transaction.

// Fix:
ret = spi_sync(spi, &bus_data->msg);
bus_data->size = 0;      // always reset, even on failure
return ret;

---

🐛 Bug: adi,ad5685 missing from DT binding

Documentation/devicetree/bindings/iio/dac/adi,ad5686.yaml, commit bf9ff8d1

The of_match table in ad5686-spi.c registers "adi,ad5685" (as well as "adi,ad5685r"), but the binding schema only lists "adi,ad5685r". Any device tree using "adi,ad5685" will fail dt-validate.

Fix: add - adi,ad5685 to the compatible enum in adi,ad5686.yaml.

---

⚠️ Documentation: LDAC GPIO polarity not stated in DT bindings

adi,ad5686.yaml and adi,ad5696.yaml, commits dbd78e3a / 31ff1408

The driver uses GPIOD_OUT_HIGH as the initial LDAC state and the trigger handler's pulse sequence (set_value(0) to latch, set_value(1) to load) is only correct when the GPIO is declared active-low in the device tree. Without GPIO_ACTIVE_LOW, the LDAC pulse is inverted and DAC outputs will never be updated atomically via the triggered buffer.

The adi,ad5791.yaml binding already models this correctly with:

ldac-gpios = <&gpio_bd 18 GPIO_ACTIVE_LOW>;

Please add a note to the ldac-gpios description in both bindings:

ldac-gpios:
  description:
    LDAC pin to be used as a hardware trigger to update the DAC channels.
    The LDAC pin is active-low; the GPIO must be declared with
    GPIO_ACTIVE_LOW polarity in the device tree.
  maxItems: 1

---

Style (checkpatch)

All 14 code commits produce:

• WARNING: Do not use whitespace before Signed-off-by: — extra blank line before each Signed-off-by: tag.
• Several commits exceed 75 characters in the description body.

These should be fixed before the series is submitted upstream.

---

Three fixup patches (using --fixup) addressing the two bugs and the documentation gap have been generated and are available in the CI artifacts.