SPI on Raspberry B+

[Bernd Walter]

On Wed, Aug 03, 2016 at 05:48:51AM +0200, Bernd Walter wrote:
> On Wed, Aug 03, 2016 at 05:28:30AM +0200, Bernd Walter wrote:
> > I'm currently writing a kernel driver for APA102 LEDs on 11-BETA3.
> > With some help of older threads on this list I've managed to
> > get my driver attached and sending data.
> > However, what I didn't manged yet is setting the SPI parameters.
> > 
> > The SPI mode seems to match fine.
> > 
> > The byte order is unfortunately wrong.
> > I need MSB first and the driver is sending LSB first.
> > Of course I could rotate all the bits, but would like to avoid it
> > for performance reasons.
> > 
> > The clock is only 500kHz according to scope.
> > I've heared that there is no device specific speed support, but since I'm
> > not using any other SPI device setting the global speed would be fine.
> 
> I've found out about sysctl dev.spi.0.clock.
> That works fine.
> 
> > I'm calling SPIBUS_TRANSFER with multiple bytes.
> > There is a gap of 2 microseconds between the bytes.
> > Don't know if it just skipping a full clock cycle and the gap will
> > be shorter with a higher clockrate, or if the driver really needs
> > 2 full microseconds between each bytes.
> > I somehow believe (and hope) it is just a full cycle in the controller.
> 
> It also remains at one clock cycle gap, so with higher speed the
> gap will be shorter as well.
> 
> Unfortunately there is no byteorder sysctl.

Oh - stupid me.
I've left some test pattern in my testcode.
The drivers byteorder is in fact MSB first.

-- 
B.Walter <bernd at bwct.de> http://www.bwct.de
Modbus/TCP Ethernet I/O Baugruppen, ARM basierte FreeBSD Rechner uvm.