* Adding Helix testing infrastructure

* Update arcade and address feedback

* Fixing tests and cancellation token source issue

* Exposed more items in IGpioController

* Updated with exposed IGpioController items

* Changed to no-op for now

Updates for supporting 64 bit linux pin nimbering in Sysfs driver

* Changes in order to make windows implementation Api compatible with Linux

* Throwing PNSE on static Create method of the UnixDriver

* Native: Libgpiod: Add GetNumberOfChips method

Some SoCs have more than one GPIO bank. For these we gave more than one entry in /dev/:
/dev/gpiochip0, /dev/gpiochip1 ... /dev/gpiochipX

This patch add a method to return how many GPIO banks the hardware have.
Signed-off-by: NMatheus Castello <matheus@castello.eng.br>

* System.Device.Gpio: GpioController: Add SetDefaultGpioChip

Some SoCs have more than one GPIO bank. For these we gave more than one entry in /dev/:
/dev/gpiochip0, /dev/gpiochip1 ... /dev/gpiochipX

This patch add a method to select the default GPIO bank to be used on the GpioController.
If the hardware has only one GPIO bank (/dev/gpiochip0 like the rpi) it will be automatically set in
the GpioController constructor. But if the hardware has more than one GPIO bank the bank to be used
for GpioController operations should be defined.

Ex: Access GPIO bank 3 (/dev/gpiochip3) line 17:

GpioController gpioCtrl = new GpioController();
gpioCtrl.SetDefaultGpioChip(3);
gpioCtrl.OpenPin(17, PinMode.Output);
Signed-off-by: NMatheus Castello <matheus@castello.eng.br>

* System.Device.Gpio: LibGpioDriver: Implement SetDefaultGpioChip

Some SoCs have more than one GPIO bank. For these we gave more than one entry in /dev/:
/dev/gpiochip0, /dev/gpiochip1 ... /dev/gpiochipX

This patch implement a method to select the default GPIO bank to be used on the GpioController.
If the hardware has only one GPIO bank (/dev/gpiochip0 like the rpi) it will be automatically set in
the GpioController constructor. But if the hardware has more than one GPIO bank the bank to be used
for GpioController operations should be defined.

Ex: Access GPIO bank 3 (/dev/gpiochip3) line 17:

GpioController gpioCtrl = new GpioController();
gpioCtrl.SetDefaultGpioChip(3);
gpioCtrl.OpenPin(17, PinMode.Output);
Signed-off-by: NMatheus Castello <matheus@castello.eng.br>

* PR #346: Review: Fix Code Style indentation

.net developers use four spaces (no tabs!)
Signed-off-by: NMatheus Castello <matheus.castello@toradex.com>

* PR #346: Review: Fix some typos and names
Signed-off-by: NMatheus Castello <matheus.castello@toradex.com>

* PR #346: Review: Get gpioChip handle by number

Use gpiod_chip_open_by_number instead the iterator way
Signed-off-by: NMatheus Castello <matheus@castello.eng.br>

* PR #346: Review: Use driver constructor instead SetDefaultGpioChip
Signed-off-by: NMatheus Castello <matheus@castello.eng.br>

* PR #346: Review: Fix build remove _gpioChipIsSet
Signed-off-by: NMatheus Castello <matheus@castello.eng.br>

* PR #346: Review: Remove GetNumberOfChips
Signed-off-by: NMatheus Castello <matheus@castello.eng.br>

* Update SysFs naming

* Trying to trick file names

* Added the file names back

* add support for spi data flow and chip select active state on linux

* add description for data flow values

* apply feedback

* Expose Raspberry register addresses

This is need for the perf critical scenarios

* address feedback

* Mark the properties as not Virtual to ensure inlining

* Dispose CancellationTokenSource

* Await task

Otherwise the tokenSource may get disposed prematurely

* Updated parameter name and summary

* Updated naming for pairs

* Added package back now that IGpioController is public

* Few code cleans

* Missed a few packages to add back

* Updated project path

* Removed packages and updated for ReadOnlySpan

* Removed unused assignments

* Added core package back

* Changed a couple readonly

* Made csproj similar to another test project

Matching existing .NET API conventions.

* Add more Mcp23xxx tests.

- Align Mcp address validation with datasheets
- Allow passing in IGpioController for master controller
- Add basic tests for reset functionality
- Fix reset caching issues
- Add validation for disabled state and tests
- Remove dead code
- Add deconstructor and bit mask helper for PinValuePair

* Fix casing to work on Unix.

* Remove tuple from the public interface.

Tuples aren't allowed in our public APIs. After some discussion creating a couple of PinVector structs to allow using bit vectors with pin/values. The 32 bit values can be used with BitVector32 if people care to. The framework currently doesn't have a 64 bit BitVector.

These vectors can potentially be expanded with additional API as we find common patterns.

Address some other feedback.

* Move PinVector to Common while #230 settles out

* Ugh, forgot the deletes.

* Add IGpioController.

This allows compositing gpio extenders/multiplexers or any other providers with device drivers.

* Hack to get command line build working

* Address feedback, fix the rest of the projects

* Add a few tests, tweak the projects to play nicer with VS

* Rejigger the abstraction in Mcp23xxx

Change the abstraction some and update the tests to map closer to the implementations. Fix a bug in setting single pin values.

* Change PinValue to struct

Allows easier interop with raw data and enables us to extend with additional functionality. This will avoid having to add additional APIs for convenience and repeated typing of `PinValue == PinValue.High ? 1 :  0`.

Doesn't impact existing enum usage.

Sample usage:

``` C#
    class Program
    {
        static void Main(string[] args)
        {
            PinValue value = true;
            Console.WriteLine($"True is {value}");
            TestApi(0);
            Console.WriteLine($"Equality is {PinValue.High == true}");
        }

        public static void TestApi(PinValue value)
        {
            Console.WriteLine($"PinValue is {value}");
        }
    }
```

* Ok, so you can't switch a struct. :) Bonus now that you don't have to. There is no such thing as an invalid PinValue anymore.

* Added editconfig file

* Beginning of cleaning

* Added summaries and sorted

* Added callback summaries

* Fixed ordering and switch/case based on feedback

* Added editconfig file

* Cleanup for SPI drivers

* Updated for feedback

* Added editconfig file

* Cleanup for I2C drivers

* Cleanup for I2C drivers

* Updated wording

* Updated for feedback

* Basic cleanup in GPIO files

* Basic comment cleanup in GPIO files

* Removed othing lingering words

* Another round of GPIO cleaning

* Removed a couple more words

* Added readonly and updated casing

* Better method grouping and synced constructor summaries

* implement fallback enumeration for hummingboard - where PWM friendly names are not currently set

* Basic comment cleanup in supporting SPI files

* Added wording from feedback

There was a typo, but also both states can have current depending on hardware configuration.  A Low has ability to sink current.