drivers: iio: adc : Add ad4695 support

[rbolboac]

No description provided.

[nunojsa]

I'll try to do a proper review in the next few days... One thing that I'll already ask is to rename your driver name.

Typically we don't name it with "wildcards" like 'x'. Just pick one of the parts the drivers supports (maybe the one you have) and name the driver against it.

[nunojsa]

Finally remembered about this one 😅

[RaduSabau1]

Hi @nunojsa, I made the changes requested by you although there is a chance I missed some. Anyway I will mark this PR as a draft since I tested it with the ZedBoard and appears like the buffer isn't working right, I managed to fix a part of it but still the behavior isn't quit the expected one.

[dlech]

FYI, we recently upstreamed the driver for this family of chips. https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git/tree/drivers/iio/adc/ad4695.c?h=testing

And @threexc is working on adding more features upstream.

[RaduSabau1]

v2 :

• Added driver dependencies from upstream
• Added driver support patches from upstream
• Added devicetree file for the zed project

[nunojsa]

@RaduSabau1 please make sure to take care about CI complains...

[nunojsa]

Still some stuff to be done

[dlech]

Just to compare notes, here is what I have been using for a .dts. Feel free to copy anything you think is useful.

A few things to point out:

I noticed that the rx-dmac@44a30000 node has a different clock and adi,destination-bus-width, so we should probably have a look at the HDL to see which one is right.

IIRC, by default, the COM pin is wired to GND on the eval board. We modified our board so I added a comment about this. But probably best to make the "official" .dts use the eval board defaults with a comment to explain what to change if you resolder the board like we did.

And finally, I know this isn't really how things are done in the ADI tree currently, but I named all of the power supplies after what there name is on the EVAL board schematic instead of which pin on the ADC they are connected to. Since the eval boards tend to have lots of jumpers that can be switched around to rewire things, I feel like doing it this way requires fewer changes to the .dts file if you end up rewiring things. And it makes it more obvious which supply exactly each node represents.

// SPDX-License-Identifier: GPL-2.0
/*
 * Analog Devices EVAL-AD4696FMCZ
 * https://www.analog.com/en/products/ad4696.html
 *
 * hdl_project: <ad469x/zed>
 * board_revision: <?>
 *
 * Copyright (C) 2024 Analog Devices Inc.
 */
/dts-v1/;

#include <dt-bindings/dma/axi-dmac.h>
#include <dt-bindings/gpio/gpio.h>
#include <dt-bindings/interrupt-controller/irq.h>
#include <dt-bindings/iio/adi,ad4695.h>

#include "zynq-zed.dtsi"
#include "zynq-zed-adv7511.dtsi"

/ {
	eval_u3: eval-board-u3-regulator {
		compatible = "regulator-fixed";
		regulator-name = "EVAL ADR4550 5.0V REFERENCE (U3)";
		regulator-min-microvolt = <5000000>;
		regulator-max-microvolt = <5000000>;
		regulator-always-on;
	};

	eval_u5: eval-board-u5-regulator {
		compatible = "regulator-fixed";
		regulator-name = "EVAL 12V/5V,-3.3V/-2.5V DUAL-LDO (U5)";
		regulator-always-on;
	};

	eval_u6: eval-board-u6-regulator {
		compatible = "regulator-fixed";
		regulator-name = "EVAL 5V/1.8V LDO (U6)";
		regulator-always-on;
	};

	eval_a8: eval-board-a8-voltage-divider {
		compatible = "regulator-fixed";
		regulator-name = "EVAL VREF/2 DIVIDER CIRCUIT (A8)";
		regulator-min-microvolt = <2500000>;
		regulator-max-microvolt = <2500000>;
		regulator-always-on;
	};

	signal_gen_neg: signal-generator-negative {
		compatible = "regulator-fixed";
		regulator-name = "Signal generator negative output";
		regulator-min-microvolt = <2500000>;
		regulator-max-microvolt = <2500000>;
		regulator-always-on;
	};
};

&fpga_axi {
	adc_trigger: pwm@44b00000 {
		compatible = "adi,axi-pwmgen-2.00.a";
		reg = <0x44b00000 0x1000>;
		#pwm-cells = <2>;
		clocks = <&spi_clk>;
	};

	rx_dma: rx-dmac@44a30000 {
		compatible = "adi,axi-dmac-1.00.a";
		reg = <0x44a30000 0x1000>;
		#dma-cells = <1>;
		interrupts = <0 57 IRQ_TYPE_LEVEL_HIGH>;
		clocks = <&clkc 17>;

		adi,channels {
			#size-cells = <0>;
			#address-cells = <1>;

			dma-channel@0 {
				reg = <0>;
				adi,source-bus-width = <32>;
				adi,source-bus-type = <AXI_DMAC_BUS_TYPE_AXI_STREAM>;
				adi,destination-bus-width = <64>;
				adi,destination-bus-type = <AXI_DMAC_BUS_TYPE_AXI_MM>;
			};
		};
	};

	spi_clk: clock-controller@44a70000 {
		compatible = "adi,axi-clkgen-2.00.a";
		reg = <0x44a70000 0x1000>;
		#clock-cells = <0>;
		clocks = <&clkc 15>;
		clock-names = "s_axi_aclk", "clkin1";
		clock-output-names = "spi_clk";

		/* Set to 2x max SCLK. */
		assigned-clocks = <&spi_clk>;
		assigned-clock-rates = <160000000>;
	};

	axi_spi_engine_0: spi@44a00000 {
		compatible = "adi,axi-spi-engine-1.00.a";
		reg = <0x44a00000 0x1000>;
		interrupt-parent = <&intc>;
		interrupts = <0 56 IRQ_TYPE_LEVEL_HIGH>;
		clocks = <&clkc 15>, <&spi_clk>;
		clock-names = "s_axi_aclk", "spi_clk";

		#address-cells = <1>;
		#size-cells = <0>;

		ad4696: adc@0 {
			compatible = "adi,ad4696";
			reg = <0>;
			spi-max-frequency = <80000000>;
			spi-cpha;
			spi-cpol;
			avdd-supply = <&eval_u5>;
			ldo-in-supply = <&eval_u5>;
			vio-supply = <&eval_u6>;
			ref-supply = <&eval_u3>;
			/* Not default: requires changing JP6 to A */
			com-supply = <&eval_a8>;
			in1-supply = <&signal_gen_neg>;
			reset-gpios = <&gpio0 86 GPIO_ACTIVE_LOW>;

			#address-cells = <1>;
			#size-cells = <0>;

			channel@0 {
				reg = <0>;
				common-mode-channel = <1>;
				bipolar;
			};

			channel@3 {
				reg = <3>;
				common-mode-channel = <AD4695_COMMON_MODE_COM>;
				bipolar;
			};
		};
	};
};

[dlech]

It looks like there are some more dependencies to cherry-pick.

fd7179e, 7758562

These go before "iio: add support for multiple scan types per channel".

And we also need to pick up 257b233 which goes right after "regulator: devres: add API for reference voltage supplies".

Are you able to compile the kernel yourself to catch these compile errors?