Re: Submit of a driver for Pi433 - a radio module for Raspberry Pi

From: Greg KH
Date: Wed Jul 12 2017 - 03:43:19 EST

On Tue, Jul 11, 2017 at 06:02:09PM +0200, Marcus Wolf wrote:
> Hi folks,
> I developed a radio shield for the 433MHz ISM band [0] for the Raspberry Pi.
> This shield is called Pi433 [1]. It can be used to communicate between two
> Raspberries or to control third party equipment e. g. cheap radio sockets [2].
> The base of this radio shield is a radio module from HopeRf - a rfm69cw-433s2.
> In the beginning, I conrolled the rf69 chip by direct access via SPI, but I
> figured out, that several tasks were hard to handle. Especially the following
> points were annoying:
> * Just one process could open the SPI interface so just one process was
> able to access the module
> * A simultaneous send and receive was very hard to implement
> Therefore I started to implement a driver.
> The drivers features:
> * simple access/nice abstraction of spi registers via open, close,
> read, write and ioctl
> * multiple applications can open
> * tx requests are queued, so every app can send a tx request at any time
> * first app asking for rx will block other apps, also asking for rx until one
> rx
> cycle is complete
> * rx cycle means listening on the air until something appears
> * if an app wants to tx during the driver is waiting in rx cycle, rx will be
> shortly interrupted to do the tx task
> * for each tx task a seperate configuration of the rf chip (bitrate,
> modulation, ...)
> can be chosen
> * wrtten in modular architecture, allowing to extend for other products with
> rf69 chip (e.g. rfm69hcw, ...) or even to support other HopeRf chips with
> similar
> interface (e. g. rfm12, rfm95, ...)
> In principle the driver was intended to be used by my project (Smarthome-Pi),
> only.
> But while having all this nice features, I was asked, whether I could offer the
> driver to the community. I never did that before, so I started reading about
> submitting drivers and figured out, that there is a lot of work to do, to meet
> all
> concepts and ideas of complying to kernel code. I spend some days and modified
> a lot, but still, a lot of things aren't perfect.
> Already known tasks are:
> * coding style does not fully comply with the kernel style guide.
> * still TODOs, annotated in the code
> * currently the code introduces new IOCTLs. I'm afraid this is a bad idea.
> Replace this with another interface, hints are welcome!
> I tested the patch on Raspbian with Kernel v4.12 and applied the patch on
> mainline Linux v4.12
> Is it posibble to integrate into the staging area to ease further development?

Sure, I'm always glad to take new drivers, but can you submit it as just
a single patch, and keep it self-contained (i.e. nothing outside of


greg k-h