On Thu, Feb 13, 2014 at 10:27:57PM +0100, Gerhard Sittig wrote:
> On Thu, Feb 13, 2014 at 16:46 +0200, Ivan T. Ivanov wrote:

> > SPI transfer lenght should be a power-of-two multiple
> > of eight bits.

> Are you suggesting that an SPI transfer cannot consist of e.g.
> three bytes?  This would be surprising, and certainly would be
> rather wrong an assumption.  Am I missing/misunderstanding
> something?

The check is supposed to be verifying that the transfer is a whole
number of the current SPI word size as I understand it (not actually
looked at the change yet but that's where it came from, factored out of
a driver).