The OMAP SPI controller drivers handle SPI_CPHA incorrectly.
It should mean:

  CPHA=0 ... sample at leading edge of clock
  CPHA=1 ... sample at trailing edge of clock

This fixes them, and the protocol drivers which depended on the wrong
meaning.  Separate ads7846 and omap_uwire patches are going upstream,
already since those drivers were already pushed there.  The tsc2102
code already had this right ... strange.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Acked-by: Kevin Hilman <khilman@mvista.com>